800D99001
 Co-Occurrence of Drinking Water
          Contaminants
Primary and Secondary Constituents
           Draft Report
        tm. S. I^AMforojioTjifMiJi^^   ^gart^
          ^


-------
                                                                                     Contents

                                           Contents

EXECUTIVE SUMMARY                                                                    ES-]
I  INTRODUCTION                                                                          I-
        1 I Purpose                                                                          1-
        1 2 Background                                                                       1-
   METHODOLOGY                                                                          2-
        2 I Database Background                                                               2-
               2 1 1  USGS NWIS Database                                                      2-
               2 1 2 Constituents of Concern                                                     2-2
               2 1 3 Threshold Values                                                          2-2
        2 2 Conventions Applied                                                                2-2
               2 2 1 Average Data for more than One Entry                                         2-2
               222 Sample Counts                                                            2-3
               223 Pearson Correlation Probability (P-value)                                       2-4
               224 Pearson Correlation Coefficient (r)                                            2-4
               225 National and Regional Analyses                                              2-5
        2 3 Methodology for Analysis                                                            2-5
               2 3 1 Statistical Analysis for Observation Points                                      2-8
               232 Frequency Counts for Observation Stations                                     2-9
   DESCRIPTION OF SECONDARY CONSTITUENTS                                             3- 1
        3 1 Background                                                                       3-1
        3 2 Antimony                                                                         3- 1
        33 Asbestos                                                                          3-2
        34 Barium                                                                           3-2
        3 5 Beryllium                                                                         3-3
        3 6 Cadmium                                                                         3-4
        3 7 Chromium                                                                        3-4
        3 8 Cyanide                                                                          3-5
        39 Iron                                                                              3-6
        3 10 Manganese                                                                       3-7
        3 1 1  Mercury                                                                         3-7
        3 12 Nickel                                                                           3-8
        3 13 Selenium                                                                        3-9
        3 14 Thallium                                                                         3-9
4  RESULTS                                                                                 4-1
        4 1 Frequency Counts                                                                  4- 1
               4 1 1  Single Occurrence Frequency Counts                                          4- 1
               4 1 2 Co-occurrence Frequency Counts (observations)                                 4-2
               4 1 3 Co-occurrence Frequency Counts (Stations)                                     4-3
        4 2 Constituent Occurrence with Depth                                                     4-4
        4 3 Ancillary Parameters                                                                4-5
        4 4 Correlation Analysis                                 . . A — rt x , ,   „    ,    ...       4-6
               4 4 1  Correlation Coefficient Calculations            U S EP* Hf ^      Llbrafy   4-6
                                                                    MaA   de
               4 42  Screening of Results                         ^ ^ ft                   _  4-7
5  SUMMARY AND CONCLUSIONS                             1 200 Pennsylvania Avenue NW  5.,
6  REFERENCES                                                 Washington DC 20460     6.,
    21, 1999                                    i                                     Draft Report

-------
                                            '''        J.
                                        	Appendices

                                          List of Appendices

                                          !  I appendices
                                            Cl appendixes
                                          "=1 C~l appendix_b
                                                ^3 observations
                                                fl stations
                                            Q| appendix_c
                                            I'l appendix_d
                                          3 LU appendix_e
                                                (" J overall
                                                Q3 sigraficant
                                                CJ threshold
                                          3 C3 appendixj
                                                _1 overall
                                                CD significant
                                             S 13 threshold
Contents/Directory Structure of CD-ROM



Directory Contents                       ^
 appendix a    Single constituent occurrence counts tabulated for ground water, surface water, and combined

 appendixjb    Constituent co-occurrence frequency tables based on observations and stations  Tables compiled
                separately for ground and surface water

 appendix_c    Well depth frequency tables for ground water samples

 appendix_d    Summary statistics for eight ancillary variables turbidity, conductance, dissolved oxygen, pH,
                hardness, alkalinity, well depth, and depth below land  Initial analysis to characterize geo-
                chemical variability by EPA Regions

 appendix e    National Level Results

                Overall Correlation coefficient results for ground and surface water Data are summarized
                irrespective of constituent threshold level

                Threshold Correlation coefficient results for ground and surface water respectively  These files
                show the correlation analyses and summary statistics for each constituent combination, including
                by threshold, and also indicate whether or not there is significance based on P-value

                Significant List of significant (based on P-value) correlation coefficient results
May 21,1999
Draft Report

-------
                                                                                            Appendices
  appendix_f    Regional Level Results
                Overall: Correlation coefficient results for ground and surface water Data are summarized
                irrespective of constituent threshold level

                Threshold  Correlation coefficient results for ground and surface water respectively These
                files show the correlation analyses and summary statistics for each constituent combination,
                including by threshold, and also indicate whether or not there is significance based on P-vaiue

                Significant. List of significant (based on P-value) correlation coefficient results
Exhibit ES-1
Exhibit 2-1
Exhibit 2-2
Exhibit 2-3
Exhibit 2-4
Exhibit 2-5
Exhibit 4-1
Exhibit 4-2
Exhibit 4-3
Exhibit 4-4
Exhibit 4-5
Exhibit 4-6
Exhibit 4-7
Exhibit 4-8
Exhibit 4-9
Exhibit 5-1
Exhibit 5-2
                                  Exhibits

Constituent pairs which co-occur on a statistically significant basis in all 10 EPA Regions      ES-3
Primary Constituent Threshold Levels                                                       2-2
Secondary Constituent Threshold Levels                                                    2-3
Critical Value of r (95% confidence level)                                                   2-5
EPA Regions Used to Partition the Data Geographically                                      2-6
Co-occurrence Analysis Flow Chart                                                        2-7
Example Single Constituent Counts for Arsenic Threshold Levels (Surface and Ground Water)   4-1
Example of Co-occurrence Frequency Counts (by observations) for Arsenic and Sulfate         4-2
Example of Co-occurrence Frequency Counts (by stations) for Arsenic and Sulfate              4-3
Comparison of Co-occurrence Frequency Percentage Based on Observations and Stations        4-3
Constituent Occurrence with Depth (percentage of observations)                              4-4
Sample Page from Appendix E - overall  National Correlation Analysis Results Ground Water   4-8
Sample Page from Appendix E - threshold  National Correlation Analysis  Correlation Coefficients
and summary statistics by threshold level (ground water)                                     4-9
Sample Page from Appendix E - significant  National correlation analysis Significant correlation
coefficients for ground water samples                                                     4-10
Statistically Significant Co-occurring Pairs                                                4-11
Constituent pairs which co-occur on a statistically significant basis in all 10 EPA Regions        5-2
Distribution of Stations by Well Depth (feet)                                                 5-2
May 121, 1999
                                      in
Draft Report

-------
                                                                   Executive Summary
EXECUTIVE SUMMARY

The purpose of this analysis is to determine whether specific primary (arsenic, sulfate, uranium,
radium, radon, nitrate) and secondary (antimony, barium, beryllium, cadmium, chromium,
cyanide, iron, manganese, mercury, nickel, nitrite, selenium, thallium, hardness, total dissolved
solids) drinking water contaminants co-occur on a statistically significant basis or if
the co-occurrence is purely a random phenomenon The information on co-occurrence will be
used by the Environmental Protection Agency (EPA) to determine the level of overlap in
regulatory requirements, for example, in cases where treatment technologies applied for one
regulation may resolve monitoring and/or additional treatment needs for another regulation  This
information may also be used to show where specific levels of one contaminant may interfere
with the treatment technology for another

This study constitutes a follow-on effort to work performed in 1994 under Work Assignment
(W A) 5 of EPA Contract 68-C3-0365 and 1998 under WA 0-25 of EPA Contract 68-C6-0059
Under the 1994 WA, SAIC supported EPA in analyzing the distribution of arsenic co-occurrence
with other inorganic constituents Under the 1998 WA, SAIC analyzed three water-quality
databases for use in co-occurrence analysis and performed initial statistical analyses for the
primary constituents (arsenic, sulfate, uranium, radium, radon, and nitrate) of concern
Preliminary and final draft reports have been previously submitted to EPA documenting this
work. In the follow-on analysis, documented in this report, the focus of the effort was on
identifying, and locating actual co-occurrence data for the primary and secondary constituents
and the determination of significant correlations of co-occurrence  In addition, co-occurrence
frequency counts were made to determine the abundance and distribution of co-occurring
constituents at specific threshold levels

As documented in the final draft report, "Co-occurrence of drinking water contaminants"
(January 27,1999), there exists several national level water quality databases that  contain
historical contaminant occurrence data  This pnor work was focused on the examination of three
primary databases (1) extracts from the USGS NWIS databases (surface and groundwater), (2)
National Ambient Arsenic Occurrence database (consists of USGS groundwater data for arsenic
samples and other constituents), and (3) the EPA State Radon Occurrence database The
conclusion of this assessment was to focus solely on the USGS NWIS database since it contained
both ground and surface water data, it was national in scope, and provided latitude/longitude
coordinates for monitoring stations that could be used in subsequent  analyses to associate
monitoring stations with public water supply systems

The NWIS database represents raw water samples The constituents analyzed have been
partitioned into two groups (primary and secondary), based on EPA priorities and technical
direction  The primary constituents include  Arsenic, Sulfate, Radon, Radium, Uranium, and
Nitrate  The secondary constituents include  Antimony, Barium, Beryllium, Cadmium,
May, 21, 1999                                ES-1                                Draft Report

-------
                                                                      Executive Summary
Chromium, Cyanide, Iron, Manganese, Mercury, Nickel, Nitrite, Selenium, Thallium, hardness
and total dissolved solids  An additional set of ancillary parameters includes  Turbidity,
Conductance, Dissolved Oxygen, pH, Alkalinity, Well Depth, and Depth Below Land  These
ancillary parameters were selected for use as indicators of hydrogeologic and geochemical
conditions that could influence the co-occurrence of specific constituents  A brief discussion on
the geochemistry, nature of release, fate and transport, and treatment for each of the secondary
constituents is also discussed in the report

The discussion of results is organized as follows single occurrence frequency counts, co-
occurrence frequency counts, constituent occurrence with well depth, regional variation in
ancillary parameters, and correlation coefficient calculations, on a national and regional basis
Three main analyses were performed,

    •   Statistically significant correlation (national and regional level)
    •   Single and co-occurrence counts
    •   Constituent co-occurrence with well depth

Correlation coefficients were calculated for all combinations of primary and secondary
constituents at all threshold levels  Calculation of the correlation coefficient and associated P-
value was used as a screening technique to filter out constituents which do not co-occur on a
statistically significant basis Calculations were made separately for ground water and surface
water on both a national and regional basis  These national level results serve as a first order
indicator of co-occurrence potential  The national level analysis acts as initial screening
technique It is first cut at identifying non-random patterns of co-occurrence Regional analysis
further refines the patterns at the national level by identifying specific geographic regions of co-
occurrence  Any pair of constituents with P-value greater than 0 05 was not considered for
further analysis
                      -—-

The analysis shows the following conclusions,

    •   For this project we analyzed a total of 22 constituents (7 primary and 15 secondary)
       Several constituents had both total and dissolved forms  In addition, each constituent was
       analyzed at one or more threshold levels  Based on these numbers, a total of 9,912 co-
       occurrence combinations were analyzed on a national level  The screening techniques
       applied to the correlation coefficients resulted in 1,860 statistically significant pairs
       (based on P-value < 0 05)   The screening results for the regional analysis showed the
       following statistics 98,185 combinations, 7,152 statistically significant pairs based on P-
       value
May 21,1999                                ES-2                                 Draft Report

-------
                                                                   Executive Summary
       Exhibit ES-1 indicates the following patterns present at the national level and also in all
       the US EPA Regions  Groundwater and surface water data show different patterns of co-
       occurrence
     Exhibit ES-1. Constituent pairs which co-occur on a statistically significant basis in
                                   all 10 EPA Regions









Ground- Water
Cadmium - Iron
Chromium - Nickel
Iron - Manganese





Surface Water
^<-
Arsenic - Nickel
Barium - Beryllium
Barium - Chromium
Chromium - Manganese
Chromium - Nickel
Iron - Manganese
Manganese - Nickel
Nitrate - Nitrite
       Radon co-occurs with other constituents in surface water in the US EPA Regions 4, 7,9,
       10  The results showed that the following pairs co-occur with radon in surface water

       Region 4   Radon - Iron, Radon - Manganese, Radon - Sulfate
       Region 7   Radon - Nitrate, Radon - Sulfate
       Region 9   Radon - Beryllium, Radon-Selenium, and Radon - Sulfate
       Region 10 Radon - Barium, Radon- Chromium, Radon - Manganese

       The number of constituent counts in general decrease with increase m well depth

       The percentage differences between the "observation counts" and the "station counts"
       were insignificant
May 21, 1999
ES-3
Draft Report

-------
                                                                        1  Introduction
1. INTRODUCTION

1.1 Purpose

The purpose of this analysis is to determine whether specific primary (arsenic, sulfate, uranium,
radium, radon, nitrate), and secondary (antimony, barium, beryllium, cadmium, chromium,
cyanide, iron, manganese, mercury, nickel, nitrite, selenium, thallium, hardness, total dissolved
solids), drinking water contaminants co-occur on a statistically significant basis or if
the co-occurrence is purely a random phenomenon The information on co-occurrence will be
used by the Environmental Protection Agency (EPA) to determine the level of overlap in
regulatory requirements, for example, in cases where treatment technologies applied for one
regulation may resolve monitoring and/or additional treatment needs for another regulation or
where water supplies may incur costs for installing multiple treatments to address other co-
occumng substances  This information may also be used to show where specific levels of one
contaminant may interfere with the treatment technology for another

1.2 Background

The 1996 Safe Drinking Water Act Amendments directed EPA's Office of Ground Water and
Drinking Water (OGWDW) to meet several regulatory initiatives with near-term deadlines  This
included final regulations for the Interim Enhanced Surface Water Treatment Rule
(December 1998), Stage I Disinfectants and Disinfection Byproducts Rule (December 1998), and
Ground Water Disinfection Rule (August 1999), and proposed regulations for Radon (August
1999) and Arsenic (January 1,2000)  EPA was already working from mandates from the 1986
amendments to regulate uranium and revise regulations for radium, gross alpha emitters, beta
emitters, and photon emitters
EPA must consider cost when establishing national primary drinking water regulations, and the
amendments added several major stipulations  First, EPA must compare the reduction m health
risks (benefits) against the cost of implementing the regulation  Second, EPA must consider any
increase in health risks that occur in complying because of changes m co-occurring
contaminants Third, EPA must account for benefits of reducing co-occurring contaminants due
to compliance with the new Maximum Contaminant Level (MCL)  Finally, EPA must eliminate
double counting benefits that result from compliance with other proposed or final regulations
A National Primary Drinking Water Regulation (NPDWR) is an enforceable standard that
controls the level of a specific contaminant that can adversely affect health and is known or
anticipated to occur in drinking water  In order to achieve this standard, EPA first sets a non-
enforceable Maximum Contaminant Level Goal (MCLG) at a level that protects against health
risks  Then EPA establishes an MCL or a Treatment Technique under the NPDWR for a
contaminant A Treatment Technique can be established in lieu of an MCL for contaminants that

May 21, 1999                                ~l                                 Draft Report

-------
                                                                        1  Introduction
cannot be measured accurately at levels of public health concern  EPA also develops
requirements for monitoring, treatment technologies, and analytical methods for the contaminant
aspartoftheNPDWR

In denvmg MCLs or treatment techniques, EPA evaluates the availability and performance of
various technologies for removing the contaminant, the ability of laboratories to measure the
contaminant accurately and consistently, the performance reliability of available analytical
methods, and the health risks associated with various contaminant levels  These analyses require
study of the occurrence of the contaminant in systems, the size of the systems, the range of the
concentrations of the contaminant in systems, and the cost of adding treatment or other measures
to achieve the MCL

Contaminant occurrence data are available from several sources  The U S Geological Survey
(USGS) has extensive data from raw surface and ground water, along with documented
analytical methodologies  These data are maintained in the National Stream Quality Accounting
Network (NASQAN) and the National Water Information System (NWIS)  EPA maintains the
Storage and Retrieval System (STORET), which contains data from States, EPA, and other
government agencies EPA conducted the Rural Water Survey between 1978 and 1980, covering
648 supplies, 474 of which were ground water supplies  The National Organics Monitoring
Survey (NOMS) was conducted by EPA from  1976 to 1977 and provides information on treated
water data from 1976 and 1977 The National  Inorganics and Radionuchdes Survey (NIRS),
conducted by EPA m 1984-1986, consists of finished drinking water samples  In addition,
regulated contaminants are tracked  for compliance purposes in the Safe Drinking Water System
(SDWIS), indicating systems that exceed MCLs, Finally, States and other groups keep
monitoring data  Quality of the data, detection levels, and reasons for sampling vary extensively
among the databases  No single database provides all the data needed for regulatory analysis of
co-occurrence of primary constituents

This study constitutes a follow-on effort to work performed in 1994 under Work Assignment
(WA) 5 of EPA Contract 68-C3-0365 and 1998 under WA 0-25 of EPA Contract 68-C6-0059
Under this 1994 WA, SAIC supported EPA m analyzing the distribution of arsenic co-occurrence
with other inorganic constituents Under the 1998 WA, SAIC analyzed three water-quality
databases for use in co-occurrence analysis and performed initial statistical analyses for the
primary constituents (arsenic, sulfate, uranium, radium, radon, and nitrate) of concern
Preliminary and final draft reports have been previously submitted to EPA documenting this
work In the follow-on analysis, documented in this report, the focus of the effort was on
identifying, and locating actual co-occurrence data for the primary and secondary constituents
and the determination of significant correlations of co-occurrence
May 21, 1999                                1-2                                 Draft Report

-------
                                                                      2  Methodology
2.
2J
METHODOLOGY
 Database Background
As documented in the final draft report, "Co-occurrence of Drinking Water Contaminants"
(January 27,1999), under EPA contract Number 68-C6-0059, Work Assignment Number 1-25,
there exists several national level water quality databases that contain histoncal contaminant
occurrence data Prior work, documented m the above mentioned report, focused on the
examination of three primary databases (1) extracts from the USGS NWIS databases (surface
and ground water), (2) National Ambient Arsenic Occurrence database (consists of USGS ground
water data for arsenic samples and other constituents), and (3) the EPA State Radon Occurrence
database
An assessment of each of these databases was previously conducted and reported on in the
January, 1997 report. The conclusion of this assessment was to focus solely on the USGS NWIS
database since it contained both ground and surface water data, it was national in scope, and
provided latitude/longitude coordinates for monitoring stations that could be used in subsequent
analyses to associate monitoring stations with public water supply systems  The NWIS data
represents raw water samples across the U S
As mentioned in the introduction, there are several other occurrence databases, such as NIRS,
NQMS, and STORET, which contain historical momtonng data  NIRS and NOMS databases
were not used due to their limited sample size, limited ability to perform specific geographic
analyses, lack of data currency, and lack of metadata  STORET data was not used because of the
lack of metadata, variability in analytical methods used across agencies contributing data, and the
decrease in the frequency of USGS updates  In addition, USGS data were obtained directly, thus
there was no need to extract it from STORET
211  USGS NWIS Database
The NWIS contains a water quality data storage retrieval system developed by the USGS Water
Resources Division  NWIS is a distributed water database in which data can be processed over a
network of computers at USGS offices throughout the U S  The system comprises the
Automated Data Processing System, the Ground-Water Site Inventory System, the Water-Quality
System, and the Water-Use Data System Within NWIS, a logical water-quality database
consists of a water-quality file (WQFILE), a station file (SITEFILE), and shared parameter
reference files (USGS,  1997a)  These three files were delivered, in ASCII format, by the USGS,
for! each of the fifty states  For each state, a set of NWIS ASCII files (WQFILE, SITEFILE, and
parameter reference file) were processed through three Statistical Analysis Systems (SAS)
programs to convert the ASCII files to dBase files based on the file layout listed in the January
report
May2J, 1999
                                      2-1
Draft Report

-------
                                                                       2  Methodology
212 Constituents of Concern

The constituents analyzed have been partitioned into two groups (primary and secondary), based
on EPA priorities  The primary constituents include Arsenic, Sulfate, Radon, Radium,
Uranium, and Nitrate  The secondary constituents include1  Antimony, Barium, Beryllium,
Cadmium, Chromium, Cyanide, Iron, Manganese, Mercury, Nickel, Nitrite, Selenium, Thallium,
hardness, and total dissolved solids  An additional set of ancillary parameters includes
Turbidity, Conductance, Dissolved Oxygen, pH, Alkalinity, Well Depth, Depth Below Land
These ancillary parameters were selected for use as indicators of hydrogeologic and geochemical
conditions that could influence the co-occurrence of specific constituents

273 Threshold Values

Based on technical direction provided by the EPA, the following threshold values were used in
the statistical analysis of co-occurrence (see exhibits 2-1 and 2-2)  For example, statistical
analysis for arsenic dissolved were performed for concentrations of 0-2,2-5, 5-10,10-20, and
>20 uG/L

                   Exhibit 2-1. Primary Constituent Threshold Levels
. ' Constituent
Arsenic - dissolved, total
Sulfate - dissolved
Radium -226
Radon - 222
Nitrate - N
Uranium 235
Uranium 238
^ / Threshol^V-*:,
2,5,10,20
25, 120, 150,250,500
3,4
100,300,500, 1000 3000
8 to
5 20, 50 60
5, 20, 50, 60
, el/nits ;4
UG/L
MG/L
PCi/L
PCi/L
MG/L
PCi/L
PCi/L
2.2 Conventions Applied

227 Average Data for more than One Entry

All statistics were computed on the daily mean values rather than the original raw data Daily
mean values were used as some stations reported multiple measurements for a parameter in one
day The majority of these measurements were taken hours apart and sometimes minutes apart
May 21, 1999
2-2
Draft Report

-------
                                                                          2  Methodology


Exhibit 2-2. Secondary Constituent Threshold Levels

Antimony - dissolved, total
Barium - dissolved, suspended, total
Beryllium - dissolved, total
Cadmium - dissolved, total
Chromium - dissolved, hexaval, total
Cyanide - total, dissolved
Iron - suspended, total, dissolved, ferrous
Manganese - suspended, total, dissolved
Mercury - dissolved, total
Nickel - dissolved, total
Nitrite - N
Selenium - total, dissolved
Thallium - dissolved, total
Hardness - total
Total Dissolved solids

6
2000
4
5
100
02
300, 2500
50
2
100
1
50
2
300
500

UG/L
UG/L
UG/L
UG/L
UG/L
MG/L
UG/L
UG/L
UG/L
UG/L
MG/L
UG/L
UG/L
MG/L
MG/L

222 Sample Counts

As a preliminary data analysis step, counts of observations (observation data and observation
stations) were made to determine the sample size of the various co-occurrence combinations The
purpose of these counts was to indicate the actual sample size available for statistical analysis
The following procedure was applied for the counts,

       Constituent 1 missing/constituent 2 present
       Constituent 1 present/ constituent 2 missing
       Constituent 1 present/ constituent 2 present

Where both constituents are present,  the observation counts were further subdivided into three
categories
       Both constituent detected
       One or the other constituents not detected
       Both constituents not detected
May 21, 1999
2-3
Draft Report

-------
                                                                         2  Methodology
223 Pearson Correlation Probability (P-value)

The Pearson Correlation Probability (P-value) associated with the Pearson Correlation
Coefficient (r) indicates whether the coefficient may be deemed statistically significant from
zero P-value is checked prior to accepting the coefficient value  If the P-value is less than 0 05
then the correlation coefficient value is deemed to be meaningful (see the following web site for
more information on P-value, http //www edsm ulst ac uk/mmitab/correlation html)  If the P-
value is large, it means that the correlation is essentially zero All results are reported at the 95%
confidence interval

224 Pearson Correlation Coefficient (r)

The Pearson Correlation Coefficient is a measure of the degree of hnear relationship between
two variables (Hoel,  1971) Correlation calculations enable the identification of the degree of
association between two measures  The Pearson correlation coefficient ranges from -1 to 1  -1
and 1 indicates the strongest amount of association  The sign of the correlation (+, -) defines the
direction of the relationship, either positive or negative A positive correlation coefficient means
that as the value of one variable increases, the value of the other vanable increases, a negative
sign indicates the opposite effect

The results obtained by the hypothesis can be achieved by another method too  The first step is
to determine when a value ofr is large enough, numerically, to refute the possible claim that x
and y are actually uncorrelated variables   The hypothesis of no correlation between x and y
would be accepted unless the magnitude ofr exceeded a critical value ofr Exhibit 2-3 below
shows critical values of r as a function of sample size   The values plotted in this exhibit are from
Murdoch and Barnes, 1974 The data plotted in this exhibit is the cntical value ofr at the 95%
confidence level  If the value ofr for a sample size falls below the curve then r is not reliable
Reliable values ofr he above the curve
May 21, 1999                                2-4                                 Draft Report

-------
                                                                         2  Methodology
                  Exhibit 2-3. Critical Value of r (95% confidence level)
           08


           07


           06


       |   05

       S
       e   °4
       o
       a
           03


           02


           01


           00
                    Correlation • Statistically Significant at 95% confidence level
                  Correlation Not Statistically Significant
                         200
                                      400           600

                                         Sample Size
800
1000
225 National and Regional Analyses

The results presented here and in previously delivered reports [final draft report (January 27,
1999) as well as on the CD-ROM delivered March 24, 1999] were separated into two levels,
national and regional  National level results included all data for the U S Regional level results
included separate analyses for each EPA Region (see Exhibit 2-4) This was necessary to see any
regional correlations as a result of heterogeneous hydrogeologic and geochemical conditions, and
to document the difference between observations made at the National and Regional  levels
    Methodology for Analysts

   analysis flow chart, as shown below (see Exhibit 2-5), was developed to ultimately determine
2.3

An
systems impacted by different threshold values  This flow chart acts like a sieve Data are
filtered through each step, and in the end gives number of drinking water systems impacted by a
threshold value  The results documented here pertain to steps 1 and 2  In step 1, constituent co-
occurrence is filtered based on significant correlation coefficients  If a constituent pair passes
through this filter, then a co-occurrence frequency table is generated  This table shows the
percentage of observations in various threshold categories  The stations associated with these
observations are also identified and can be mapped   This station mapping  forms the basis for
Mayl21,1999
                                           2-5
                Draft Report

-------
                                                                                           2  Methodology
step 3 which is to geographically associate Public Water Supplies with the co-occurrence data at
these stations  The methodology for step 3 is currently being discussed  This methodology to
associate PWS's with the data, was described at a project meeting held March 24,1999
              Exhibit 2-4. EPA Regions Used to Partition the Data Geographically
 Region 1 - responsible wrthm the slates of Lopnsc'i.ut  v'aiie M=)ssac"u;pt s flev Ha.rfwre F ice* bianj -aid Veirorit

 Region 2 responsible within the slates of fle\  J>-r:>s<  Je /' ci* a ij 11«  en ctp> o""'j^-o Pice sit! ihe IJ £ '-ug'j Isionco

 Region 3 - responsible within the slates of DeU \*ra  •. at taoj Psri.)va)3 viijna  V-">=l   icjna a'id  t-e 3islnct ot Sour-ibid

 Region 4 responsible within the states of A-hbsrnd  "oida Gsorq'3, ( tn^uckj "»issiC'siuoi  NotiCarolna Ccu r Ca clina ordT

 Region5 responsible wrthin the states of I'lruo  iidirt  \rlniir  Mni ^uta Ch'c  anJ V*/'^c3-ibii

 Region 6 responsible within the states of-ikar-c  LOJ .ui j Ne^Mo'ij Ci-jhoii j  aiJ TL  t.

 Remon 7 responsible within the states of io-va san^i. hii—otr  a J r ttraskj

 Region 8 responsible within the states of Cou aiJj  v'uijra  \u t» D«ko <»  Zcul,i In*o ,, Utah ardWfjnmi

 Remon 9 responsible wrthin the states of •vuotia C«nbria H-i>-n f-.t-.'^ija «riJ I >s l« i uf e^ c.1 Gut) t 4 »J Amern.* i Soriua

 Region 10 - responsible within the stales of xi 
-------
                                                                    2  Methodology
       Stepl
                    Exhibit 2-5. Co-occurrence Analysis Flow Chart
                           NO—* Reject the Constituent Pair
                             Yes
       Step 2
       Generate
'Co-occurrence frequency"
.table based on threshold,
         levels
                    Select threshold levels of interest
                     Locate observation
                   stations associated with
                  selected threshold values
                                Completed
       Step 3
         PWS
     Geographically
     Associated with
         Data
                                               To be completed
NO—*No Systems Impacted
                            Yes
                     Number of Systems
                         Impacted
May 21, 1999
                      2-7
                            Draft Report

-------
                                                                        2  Methodology
231  Statistical Analysis for Observation Points

There are approximately 500,000 data points in the NWIS database for primary and secondary
constituents This includes multiple entries for the same station As discussed below three sets of
analyses were performed for all primary and secondary constituents

•   Single constituent count: As per EPA's technical direction, single constituent counts were
performed for all primary and secondary constituents Threshold values used for these
constituents are listed in Exhibits 2-1 and 2-2   This analysis was performed as a screening level
approach to determine the abundance and distribution of constituents prior to performing co-
occurrence counts as described below  It gives information about the data available above or
below a threshold level for that constituent  Examination of this data provides an initial
screening technique to reduce the number of co-occurrence tables to be generated later

•   Scatter Plot: Scatter plots were generated for each constituent These plots were generated to
discern whether there is any pronounced relationship, and if so, whether the relationship is linear
These plots were delivered previously at the March 24, 1999 project meeting

•   Calculation of Correlation Coefficient and P-value: Pearson Correlation Coefficients and
associated P-values for combinations of primary and secondary constituents were calculated
These analyses were done separately for ground water and surface water on both a national and
regional level

•   Co-occurrence frequency counts  Co-occurrence counts of the primary constituents with
other selected constituents were performed as per EPA's technical direction  This analysis was
performed to see how much data lies above or below a particular range of values for the selected
constituents The ranges of values used for arsenic are as follows

-   Arsenic (ND, D-2,2-5, 5-10, 10-20, >20ug/l)
-   Arsenic (as above)/Sulfate (ND, -25,25-120, 120-250,250-500, >500 MG/1)
-   Arsenic (as above)/Uramum (ND, D-5, 5-20,20-50, 50-60, >60 Pci/1)
-   Arsenic (as above)/Radium-226 (ND, D-2,2-3, 3-4,4-5, >5Pci/l)
-   Arsenic (as above)/Radon (ND, D-100,100-300, 300-1000, 1000-3000, >3000Pci/l)
-   Arsenic (as above)/Nitrate (ND, D-5, 5-8, 8-10, > 10 MG/1)
-   Arsenic (as above)/Iron (ND, D- 3, 3-1  5, 1 5-2 5, >2 5 MG/1)
-   Arsenic (as above)/Manganese (ND, D-20,  20-50, >50ug/l)
-   Arsenic (as above)/Hardness (ND, D-300, >300 MG/1)
-   Arsenic (as above)/TDS (ND, D-500, >500 MG/1)
-   Arsenic (as above)/Nitnte (ND, D-1, > 1 MG/1)
-   Arsenic (as above)/Antimony (ND, D- 006, > 006 MG/1)
-   Arsenic (as above)/Barium (ND, D-2, >2MG/1)

May 21, 1999                                2^8                                 Draft Report

-------
                                                                       2  Methodology
-   Arsenic (as above)/Berylhum (ND, D- 004, > 004MG/1)
-   Arsenic (as above)/Cadmium (ND, D- 005, > 005 MG/1)
-   Arsenic (as above)/Chromium (ND, D-1, > 1 MG/1)
-   Arsenic (as above)/Cyanide (ND, D- 2, > 2MG/1)
-   Arsenic (as above)/Mercury (ND, D- 002, > 002 MG/1)
-   Arsenic (as above)/Nickel (ND, D- 05, > 05 MG/1)
-   Arsenic (as above)/Selemum (ND, D- 05, > 05 MG/1)
-   Arsenic (as above)/Thalhum (ND, D- 002, > 002 MG/1)

Note  ND = non-detect (value below detection level), D - detect (value above detection level)

In addition, a set of co-occurrence counts was created (based on EPA technical direction) for the
following sets of constituent pairs using the threshold values indicated above,

SO4 - Uranium, Radium, Radon, Nitrate, Iron, Manganese, Hardness, Total Dissolved Solids,
Nitrite.

Uranium - Radium, Radon, Nitrate, Iron, Manganese, Hardness, Total Dissolved Solids, Nitrite

Radium - Radon, Nitrate, Iron, Manganese, Hardness, Total Dissolved Solids, Nitrite

Radon - Nitrate, Iron, Manganese, Hardness, Total Dissolved Solids, Nitrite

These tables of frequency counts were made separately for ground and surface water

232  Frequency Counts for Observation Stations

There are approximately 40,000 observation stations in the NWIS database  A similar analysis
was performed to tabulate co-occurrence frequencies based on observation stations For
example, instead of counting multiple hits at a single station, each station was counted as a single
hit if multiple observations were made

Co-occurrence counts for arsenic, sulfate  uranium, radium, and radon in ground water stations
were initially generated to compare to the observation point results
May 21 1999                                2-9                                Draft Report

-------
                                              3  Description of Secondary Constituents
3. DESCRIPTION OF SECONDARY CONSTITUENTS

3.1 Background

This chapter presents a brief hydrogeochemical discussion of the secondary constituents  The
purpose of this discussion is to present factors affecting fate and transport of the constituents
This discussion includes geochemistry, nature of release, fate and transport, and treatment for
each of the secondary constituents The primary constituents were previously discussed in the
January 27, 1999 report  Asbestos, a secondary constituent originally specified by EPA, is
included in the discussion however no data was available for statistical analysis in the NWIS data
base

With the exception of hardness, total dissolved solids, nitnte-N, and cyanide, the secondary
constituents are comprised mainly of metals Metals have fairly limited mobility in soil and
ground water because of cation exchange or sorption on the surface of mineral grains They can
also form precipitates of varying solubility under specific Eh-pH conditions The pH descibes the
number of protons and the Eh is related to the number of electrons  If a solution has several ions
present that can react to form different products or occur in different valence states, the stable
product or valence state at a given concentration of reactants will be a function of the pH and Eh
of the solution (Fetter, 1993) Metals are mobile in ground water if the Eh-pH range is such that
soluble ions exist and the soil has a low cation-exchange capacity They can also be mobile if
they are chelated or if they are attached to a mobile colloid Conditions that promote mobility
include an acidic, sandy soil with low organic and clay content  Discharge of a metal in an acidic
solution would keep the metal soluble and promote mobility (Fetter, 1993)

3.2 Antimony

Antimony is a metal found in natural deposits as ores containing other element The most
common antimony ores are the sulfide (stibnite) and the tnoxide (valentmite)  Industrial dust
and exhaust gases of cars and oil fuels are the main sources of antimony in urban air Main
sources of contaminants in drinking water are discharge from petroleum refinery, fire retardant,
ceramics, electronics, and solder  Substantial amounts of antimony tnoxide are released to the
atmosphere during processing of antimony materials including smelting of ores, molding and
incineration of products, as well as the combustion of fossil fuels which utilize the high
temperatures needed to volatilize antimony tnoxide (EPA, 1995)

Little information is available on the transformations and transport of antimony in various  media
The mobility of antimony in soils is not clearly understood The strength of its adsorption to soil
and sediments depends upon a variety of factors such as pH, organic matter content, as well as
Oxidation State of the particular ion (EPA, 1995)
May 21, 1999                                 3-1                                 Draft Report

-------
                                              3  Description of Secondary Constituents
Best available technologies for antimony treatment are coagulation/filtration and reverse osmosis
(EPA 1995)

3.3 Asbestos

Asbestos was initially included as a secondary constituent, however, the USGS NWIS extract did
not contain any asbestos records  Asbestos is a fibrous material occumng in natural deposits It
is composed of silicon, oxygen, hydrogen, and metal cations (sodium, magnesium, calcium, and
iron) (Dojhdo and Best,  1993)

Asbestos can be introduced into natural waters by erosion of asbestos containing minerals and
ores Natural weathenng by wind and water and mining activities can lead to dispersion of
asbestos into rivers and lakes However, the principal source of asbestos in some water bodies is
the disposal of industrial wastes (Toft et al, 1984) Asbestos fibers have been released into water
by the dumping of mining tailings into lakes, by the runoff of process and air scrubber water into
lakes and streams, and by the use of asbestos cement pipes in water supply systems The largest
releases of asbestos occurred in Pennsylvania and Louisiana (US EPA, 1995)

Because asbestos fibers in water don't evaporate into air or breakdown in water, small fibers and
fiber-containing particles may be carried long distances by water currents before settling to the
bottom, large fibers and  particles tend to settle more quickly  Asbestos does not tend to adsorb to
solids normally found in natural water systems, but some materials (trace metals and organic
compounds) have an affinity for asbestos minerals  The fibers are not able to move down through
soil to ground water Asbestos is not affected by photolytic processes and is considered to be
non-biodegradable by aquatic organisms Asbestos fibers are  not broken down to other
compounds in the environments and, therefore, can remain in the environment for decades or
longer (EPA, 1995)

Best available technologies for treatment of drinking water is coagulation/filtration, direct and
diatomite filtration, and corrosion control (EPA,  1995)

3.4 Barium

Barium exists in nature only in ores containing mixtures of elements The most common ores are
the sulfate, barite, and the carbonate, withente

Barium exists in water as Ba2+ ion  In strongly mineralized waters it forms ion pairs with
carbonates and sulfates  Barium is a common constituent of surface waters, but in low
concentrations  This is because of its low solubility and its ready adsorption onto participate and
sediments (Dojhdo and Best, 1993) Barium's distribution in  ground water is controlled by the
solubility of barite (BaSO4) By its chemical nature, it is unlikely to be found m water as a

May 21, 1999                                K2                                 Draft Report

-------
                                               3  Description of Secondary Constituents
banum ion Higher concentrations in ground water are expected to be found than in surface water
(Zuane, 1990)

Barium is released to water and soil in the discharge and disposal of drilling wastes, from the
smelting of copper, and the manufacture of motor vehicle parts and accessories Barium is
emitted into the atmosphere mainly by the industrial processes involved in the mining, refining,
and production of banum and barium-based chemicals, and as a result of combustion of coal and
oil (EPA,  1995)

Best Available Technologies for treatment of of banum in drinking water are Ion Exchange,
Reverse Osmosis, Lime Softening, Electrodialysis

3.5 Beryllium

Beryllium is concentrated in silicate minerals relative to sulfides and in feldspar minerals relative
to ferro-magnesium minerals Beryllium occurs in trace quantities in the environment, except in
beryllium ore bodies

Beryllium exists in water as Be2+, but is largely covalent in character and forms organic
complexes Beryllium metal dissolves in strong acids and strong bases (amphotenc) The
chloride and nitrate constituents are water soluble, and sulfate only partially (Dojhdo and Best,
1993) Beryllium oxide and hydroxide have low solubilities and can act as a control on beryllium
concentration (Fetter, 1993)

The solubility limit of beryllium in fresh natural waters at near neutral pH is somewhat greater
than 0 001 MG/L Beryllium oxide and hydroxide have low solubilities and can act as a control
on beryllium concentration Ground waters tend to contain very little beryllium, rarely exceeding
0 002 MG/L  Beryllium is not likely to be found in natural water above trace levels due to the
insolubility of oxides and hydroxides at the normal pH range Its environmental concentration is
further limited by the tendency of aqueous beryllium to be sorbed onto materials high cation
exchange capacities, such as clays  It  has been reported to occur m US drinking water at 0 01 to
0 7 Ug/L (US EPA, 1995)

Beryllium enters the environment principally from coal combustion Other sources of Beryllium
contaminant in drinking water are discharge from  metal refineries and coal burning factories,
discharge from electrical, aerospace, and defense industries Beryllium release in US is mostly to
land It is primarily from copper rolling and drawing industries, which use it as a hardener alloys
The largest release occurred m Pennsylvania and Ohio (US EPA,  1995)

Best available technologies for water treatment are activated alumina, coagulation/filtration, ion
exchange, lime softening, and reverse osmosis

May 21, 1999                                3~3	     :            =   ~~t

-------
                                              3  Description of Secondary Constituents
3.6 Cadmium

Cadmium does not exist in nature as thje native metal but principally as the sulfide ore
(greenockite)

Cadmium occurs in aqueous solution in the +2 valence state  The free Cd2+ ion occurs at pH
levels below 8 At higher pH values, [Cd(OH)]* is formed, and in very alkaline solutions
[Cd(OH)3] and [Cd(OH)4]2' are formed In surface waters, cadmium is adsorbed onto humic
substances and to paniculate matter, especially mineral solids As an element cadmium is
insoluble in water Chlorides, nitrates, and sulfates of cadmium are soluble in water  Cadmium
will precipitate at high pH since carbonate and hydroxide are insoluble (Zuane, 1990) Under
reducing conditions with sufficient sulfide present, Cadmium form sulfide minerals which are
generally very insoluble as long as reducing conditions are maintained (Deutsch, 1987)  Low
solubility of cadmium carbonate can act as a control on solubility of cadmium (Fetter, 1993)
Cadmium does not stay in solution for long in natural waters, either it is precipitated as carbonate
or it is adsorbed onto paniculate matter and incorporated into bottom sediments Once it has been
incorporated into the sediment, cadmium does not re-dissolve in neutral or alkaline conditions,
because in the anoxic conditions in the sediments, cadmium sulfide is formed, which is insoluble
(Dojhdo and Best, 1993)

Cadmium is introduced into the environment in the waste waters of industnes using cadmium.
but also in discharges from the iron and steel industry These deposits can serve as sources to
ground and surface waters, especially when in contact with soft, acidic waters Major industrial
releases of cadmium are due to waste streams and leaching of landfills, and from a variety of
operations that involve cadmium or zinc The remaining cadmium emissions are from fossil fuel
combustion, fertilizer application, and sewage sludge disposal Cadmium also occurs as a by-
product of corrosion of some galvanized plumbing and distribution system materials Cadmium
adsorbed to mineral surfaces (e g  clay) or organic materials would be more easily
bioaccumulated or released in the dissolved state when sediments are disturbed, such as during
flooding Cadmium is not known to form volatile compounds in the aquatic environment (US
EPA, 1995)

Best Available Technologies for water treatment are Coagulation/Filtration, Ion Exchange, Lime
Softening, and Reverse Osmosis (US EPA, 1995)

3.7 Chromium

Chromium is  a naturally occurring metal in drinking water Chromium in natural waters occurs
in a +3 and a +6 valence state  Hexavalent chromium in ground water is soluble and mobile and
trivalent chromium will be insoluble and immobile (Fetter, 1993)  In natural surface waters,
chromium occurs mainly as either Cr(III) or Cr(VI), depending on the conditions In  well aerated

May 21, 1999,                               3^4Draft Report

-------
                                              3  Description of Secondary Constituents
waters, Cr(Vl) will be the principal form and this, in anaerobic condition, is reduced to Cr(III)
which then forms colloidal hydrous oxides which in turn are adsorbedonto suspended particulate
matter The amount of reduction that takes place is dependent on the amount of organic matter
present  If this is low, Cr(VI) can be stable for long periods Cr(VI) is also readily reduced by Fe
2+ and sulfides which are often present in natural watersCr(III) is however slowly oxidised to
Cr(VI) by dissolved oxygen but more readily by MnO2 (Dojhdo and Best, 1993)

Chromium occurs in nature mostly as chrome iron ore, or chromite The two largest sources of
chromium emission in the atmosphere are from the chemical manufacturing industry and
combustion of natural gas, oil, and coal Other sources include wind transport from road dust,
cement producing plants because cement contains chromium, the wearing down of asbestos
brake linings from automobiles or similar sources of wind carried asbestos since asbestos
contains chromium, incineration of municipal refuse and sewage sludge, exhaust emission from
automotive catalytic converters, emissions from cooling towers that use chromium compounds as
rust inhibitors, waste waters from electroplating, leather tanning, and textile industries when
discharged into surface waters, and solid wastes from chemical manufacture (US EPA, 1995)

Best Available Technologies for water treatment are Coagulation/Filtration, Ion Exchange,
Reverse Osmosis, Lime Softening (for Cr(III) only) (US EPA, 1995)

3.8 Cyanide

Cyanide is a carbon-nitrogen chemical unit, which combines with many organic and inorganic
compounds  Cyanides may occur in water as HCN (hydrocyanic acid), CN- ion (simple cyanide)
or in the form of complexes The cyanide ion combines with many metal ions to form complexes
of varied stability The most stable complexes are formed with iron and cobalt  The complexes
formed with cadmium, zinc, copper, and nickel are less stable The less stable complexes readily
dissociate m water to release free cyanide ion (Dojhdo and Best, 1993)

The major sources of cyanide releases to water are reported to be discharges from metal finishing
industries, iron and steel mills, and organic chemical industries Releases to soil appear to be
pnmanly from disposal  of cyanide wastes in landfills and the use of cyanide-containing road
salts Cyanide released to air from car exhaust is expected to exist almost entirely as hydrogen
cyanide gas (US EPA, 1995)

Soluble cyanide compounds such as hydrogen and potassium cyanide have low adsorption to
soils with high pH, high carbonate and low clay content However, at pH less than 9 2, most free
cyanide is expected to convert to hydrogen cyanide, which is highly volatile Soluble cyanides
are not expected to bioconcentrate Insoluble cyanide compounds such as the copper and silver
salts may adsorb to soils and sediments, and generally have the potential to bioconcentrate
Insoluble forms do not biodegrade to hydrogen cyanide (US EPA,  1995)

May 21, 1999                                3^5                                Draft Report

-------
                                               3  Description of Secondary Constituents
Best Available Technologies for treatment are Ion Exchange, Reverse Osmosis, Chlorine (US
EPA, 1995)

3.9 Iron

Iron is a common constituent of anoxic groundwater Anoxic groundwater oxidizes dunng
drinking water production, precipitating Fe-oxyhydroxides Iron exists in water either as true
solution, as a colloid or in the form of suspended particles, depending on such environmental
factors as the concentration of organic matter, dissolved oxygen, carbon dioxide, pH, and extent
of microbial activity.

Iron is mainly present in the feme Fe3+ state and its salts are readily hydrolysed to to insoluble
forms, so the concentration of iron in water is generally low In anoxic conditions, however,
ferrous (Fe2+) ions are formed, and many of its compounds are soluble (Dojlido and Best, 1995)
Feme and ferrous compounds (eg Chlorides) are highly soluble in water, but ferrous iron is
readily oxidized to form insoluble feme hydroxides that flocculate and settle, therefore, iron is
expected in low concentrations in surface water (Zuane, 1990) Under oxidising conditions at pH
values greater than about 5 5, iron can occur in more than one valence state, form relatively
insoluble minerals Total iron solution is limited by femhydnte (Fe(OH)3) Under reducing
conditions with sulfide present iron form insoluble sulfide mineral (Deutsch, 1997)

Ferric and ferrous compounds (eg Chlorides) are highly soluble in water, but ferrous iron is
readily oxidized to form insoluble feme hydroxides that flocculate and settle, therefore, iron is
expected in low concentrations in surface water (Zuane, 1990) Under oxidising conditions at pH
values greater than about 5 5, iron can occur in more than one valence state, form relatively
insoluble minerals Total iron solution is limited by femhydnte (Fe(OH)3) Under reducing
conditions with sulfide present iron form insoluble sulfide mineral (Deutsch, 1997)

The mam sources for Fe2" in groundwater are the dissolution of Fe(ll) bearing minerals and the
reduction of Fe-hydroxides present in the sediments (Appelo and Postma, 1990) Presence of iron
in surface water is due to weathering and leaching of rocks and soils, discharges from foundanes,
and corrosion of pipes and storage tanks In ground water the high content of iron can be due to
the  frequency of the elevated iron level in the earth strata related to feeding aquifers  Acid mine
drainage (leaching) may be a source of iron contamination

The most obvious control on Fe2^ concentrations in groundwater is by oxidation to Fe3* and
precipitation as Fe-oxyhydroxides This process occurs particularly in zones of ground water
discharge, near rivers where Fe2~  rich ground water comes in contact with atmospheric oxygen
In aquifers Fe-oxyhydroxides may control the solubility of dissolved Fe2+, and the produced Fe-
oxyhydroxides may be transported along with the groundwater (Appelo and Postma, 1994)
May 21,1999                                3-6                                 Draft Report

-------
                                              3  Description of Secondary Constituents
Treatment technologies for Fe2+ are filtration, reverse osmosis, distillation, cation exchange,
mineral bed Treatment technologies for Fe3* are reverse osmosis and distillation (US EPA,
1995)

3.10 Manganese

Manganese is a naturally occumng substance It does not occur in the environment as pure metal
but occurs in combination with other chemicals (oxygen, sulfur, and chlorine) Manganese is
chemically similar to iron

Oxygenated water will have low levels of manganese In surface water manganese is trapped
within suspended organic matter particles The practical solubility limit for manganese is highly
dependent on the Eh and pH of the solution While pervasive at trace to significant
concentrations in geological matenals, manganese concentration in natural waters is highly
variable, because of the strong influence of acidity and redox conditions on manganese
solubility  Not only is its aqueous geochemistry complex because of this influence, but
manganese behavior is highly coupled to that of iron in many waters  Under oxidising conditions
at pH values greater than about 5 5, manganese can occur in more than one Valence State, form
relatively insoluble minerals  MnO2 limits manganese concentration (Deutsch, 1997)

Oxygenated water will have low levels of manganese In surface water manganese is trapped
within suspended organic matter particles The practical solubility limit for manganese is highly
dependent on the Eh and pH of the solution While pervasive at trace to significant
concentrations in geological materials, manganese concentration m natural waters is highly
variable, because of the strong influence of acidity and redox conditions on manganese
solubility  Not only is its aqueous geochemistry complex because of this influence, but
manganese behavior is highly coupled to that of iron in many waters  Under oxidising conditions
at pH values greater than about 5 5, manganese can occur in more than one Valence State, form
relatively insoluble minerals  MnO2 limits manganese concentration (Deutsch, 1997)

Iron and manganese are concentrated in water by contact with rocks and minerals, occasionally
man-made materials like iron and steel pipes  Water percolating through soil and rock can
dissolve minerals containing iron and manganese and hold them in solution

Treatment technologies for manganese are reverse osmosis, distillation, and cation exchange (US
EPA, 1995)

3.11 Mercury

Mercury is the only metal that is liquid at room temperature Mercury is found in nature as the
metal and the ore cinnabar

May 21, 1999                                 3^7                                 Draft Report

-------
                                              3  Description of Secondary Constituents
Mercury occurs as a metal and in the valence states +1 and +2  Most of the inorganic mercury
compounds have a low solubility. Under most natural conditions, there is little soluble inorganic
mercury (Fetter, 1993)

Water borne pollution may originate in sewage, metal refining operations, or most notably from
chloralkah plants  Ninety percent of the mercury was released to land in USA  These releases are
primarily from chemical and allied industries The largest releases occurred in Tennessee and
Louisiana. The largest direct releases to water occurred in West Virginia and Alabama (US EPA,
1995)

Mercury is introduced into surface waters mainly as the metal and as phenyl mercuric acetate In
water it exists as the mercuric ion Hg2+ and as elemental mercury Hg ° The form in which
mercury occurs depends upon pH, redox potential, and the type and concentration of anions
present which can react with mercury (Dojhdo and Best,  11993) Two characteristics - volatility
and biotransformation - make mercury somewhat unique as an environmental toxicant Its
volatility accounts for atmospheric concentration up to 4 times the level of contaminated soils in
an area  Inorganic forms of mercury can be converted to organic forms by microbial action in
biosphere

Best available technologies for water treatment are coagulation/filtration, granular activated
carbon,  lime softening, and reverse osmosis Coagulation/filtration, hme softening, and reverse
osmosis are recommended only if influent Hg concentrations do not exceed 10 Ug/L (US EPA,
1995)

3.12 Nickel

Nickel is found m many ores as sulfides, arsenides, antimomdes, and oxides or silicates Nickel
can occur m the oxidation states ranging from -1 to +4, but in aqueous solutions +2 valence state
(mckelous) dominates Nickelous ion forms stable complexes with both organic and inorganic
hgands and is also adsorbed onto paniculate matter The commonest inorganic ligands in natural
water are hahdes, sulfates, phosphates, carbonate and carbonyls  Humic  and fulvic acids form
medium-string complexes with nickel, and as a result nickel is  fairly mobile metal in natural
waters (Dojhdo and Best, 1993)

Nickel is most mobile of the heavy metals in aqueous environment Once nickel is m surface and
ground water systems, physical and chemical interactions (complexation,
precipitation/dissolution, adsorption/desorption, and oxidation/reduction) determines nickel's
fate (EPA, 1995)

Best available technologies for nickel treatment are ion exchange, lime softening, reverse
osmosis (US EPA, 1995)

May 2!, 1999                                 3^8                                Draft Report

-------
                                              3  Description of Secondary Constituents
3.13 Selenium

Selenium occurs occasionally as the native element, but usually in association with sulfur It is
found mainly as selemdes in ores of iron, lead, silver, and copper Selenium is regarded as a
metalloid, i e, it has properties of both metals and non-metals It forms compounds of three
valence states, Se2+, Se4+, and Se6+ (Dojlido and Best, 1993)

Selenium occurs in oxidising solutions as selenit, with a +4 valence and as selenate, with a +6
valence It can be reduced to the insoluble elemental form, Se° It may also form a precipitate
ferroselamt, FeSe2, under reducing conditions  Selanate may be sorbed onto amorphous feme
hydroxides (Fetter, 1993) Selenium concentration in fresh water is usually around 0 02 ppm
The selenium content of surface water is greatly influenced by pH, being high in acidic (pH <
3 0) and in alkaline waters (pH > 7 5) (USEPA , 1995)

Selenium is expected to be found in raw water from soil contamination in areas nch in selenium,
it appears in trace quantities only in public sewers due to industrial pollution (Zuane, 1990)
Selenium concentrates in irrigation return water draining from land that has soil high in
selenium The principal source of selenium m the environment appears to be burning of fossil
fuels and in cement production (Dojlido and Best, 1993)

Best available technologies for treatment of selenium are activated alumina
coagulation/filtration (Se/VI only), hme softening, reverse osmosis, and electrodialysis (US
EPA, 1995)

3.14 Thallium

Thallium is a rare element  It is found in potassium minerals (feldspar and mica) and also in
sulfide ores of zinc and lead Thallium exists in water as the thallous ion Tf and the thalhc ion
Ti3+, with the Ti+ being the dominant form in oxygenated water Soluble thaihc compounds form
colloidal oxides in natural waters, which then precipitate  onto sediments In the reducing
conditions m the sediments, thallous ions are formed which form insoluble Ti2S with sulfide
Thallium is readily adsorbed onto paniculate clay material in water and removed from the water
column (Dojltdo and Best, 1993)

The best available technologies for thallium m drinking water are activated alumina and ion
exchange (US EPA, 1995)
May 21,1999                                3-9                                 Draft Report

-------
                                                                             4  Results
4. RESULTS
This section presents a discussion of the statistical results for the co-occurrence of the primary
and secondary constituents  The discussion is organized as follows  frequency counts and single
occurrence frequency counts, co-occurrence frequency counts, constituent occurrence with well
depth, regional variation in ancillary parameters, and correlation coefficient calculations, on a
national and regional basis

4.1 Frequency Counts

411  Single Occurrence Frequency Counts

Single constituent counts are documented in Appendix A for combined (A-l), ground water (A-
2) and surface water (A-3) samples  Appendix A gives information about the data available
above or below a threshold level for that particular constituent  Exhibit 4-1, for example, shows
the constituent count for arsenic (combined ground water and surface water)  These tables were
generated using data on a national level
  Exhibit 4-1. Example Single Constituent Counts for Arsenic Threshold Levels (Surface
                                  and Ground Water)
Constituent
ARSENJC-D1SSOLVED-PO 1 000-(UG/L)
ARSENIC-D1SSOLVED-P01000-(UG/L)
ARSENIC-DISSOLVED-PO 1 000-(UG/L)
ARSENIC-D1SSOL VED-PO ! 000-( UG/U
ARSENIC-DISSOLVED-P01000-(UG/L)
/•
ARSENIC-DISSOLVED-PO 1 000-(UG/L)
ARSEN1C-TOTAL-P01002-(UG/L>
ARSEN1C-TOTAL-POI002-(UG/L)
ARSENIC-TOTAL-PO I002-(UG/L)
ARSENIC-TOT AL-P01002-(UG/L)
ARSENIC-TOT AL-P01002-{UG/L)
ARSENIC-TOTAL-PO 1 002-(UG/L)
Threshold
ND
D<=2
2<=->
5<=IO
I0<=20
>20
ND
D<=2
2<=5
3<=IO
I0<=20
>20
Count
8149
495S
3324
1963
1436
1641
5053
2>9I
1530
727
640
685
Percent
3796
2308
H48
9 14
669
764
4501
2308
1363
648
57
6 I
Note ND = no-detect, D = detect
May 21, 1999
4-1
Draft Report

-------
                                                                            4  Results
For dissolved arsenic, nearly 40% of the samples are below the detection limit  Less than 8% of
the samples are above 20 ug/1 dissolved arsenic Appendix A gives a complete set of tables for
the remaining primary and secondary constituents

412 Co-occurrence Frequency Counts (observations)

Co-occurrence counts of the constituent pairs were developed and are documented in Appendix
B  Appendix B shows the percentage of observations which are above specific threshold values.
An example of one of these tables in shown in Exhibit 4-2 (Co-occurrence Frequency Count for
Arsenic and Sulfate in Ground Water) The arsenic threshold values are shown along the left-
hand side of the table The sulfate threshold values are shown along the top row  For each row
in the table, the upper number is the count and the lower number is the percentage This exhibit
shows that only 2 5% of the data are above the threshold values of arsenic > 20 ug/1 and sulfate >
120 mg/1 The data are compiled separately for ground water (Appendix B-obs-gw) and surface
water (Appendix B-obs-sw)

  Exhibit 4-2. Example of Co-occurrence Frequency Counts (by observations) for Arsenic
                                      and Sulfate
Ground Water Stations
TABLE OF ARS1 BY SULF
ARS1(ARSEN1C-D1SSOLVED-P01000-(UG/L)) SULF(SULFATE-DISSOLVED-P00945-(MG/L))

ND
EK=2
?<— 5

5<=10
1 0<— 20

>20
Total
Count
%
Count
%
Count
%
Count
%
Count
%
Count
%
Count
%
ND
121
141
27
031
19
022
12
0 14
19
022
20
023
218
253
EK=25
1891
21 97
7i4
829
446
5 18
278
323
196
228
216
251
3741
4346
25<=120
916
1064
653
759
499
58
309
359
244
283
274
3 18
2895
3363
120<=250
217
252
115
134
106
1 23
75
087
87
1 01
127
148
111
845
250<=500
133
I 55
103
12
46
053
55
064
36
042
34
039
407
473
>500
349
405
79
092
67
078
37
043
35
041
53
062
620
72
Total
3627
42 14
1691
1964
1183
1374
766
89
617
7 17
724
841
8608
100
Frequency Missing = 137245
Note D = detect, ND = non-detect
May 21, 1999
4-2
Draft Report

-------
                                                                            4  Results
413 Co-occurrence Frequency Counts (Stations)

Co-occurrence frequency counts by stations (Appendix B-sta-gw) for arsenic, sulfate, uranium,
radium and radon were also performed for ground water samples An example of one of these
tables is shown in Exhibit 4-3

  Exhibit 4-3. Example of Co-occurrence Frequency Counts (by stations) for Arsenic and
                                        Sulfate
Ground Water Stations
TABLE OF ARS1 BY SULF
ARS1(ARSENIC-DISSOLVED-P01000-(UG/L)) SULF(SULFATE-DISSOLVED-P00945-(MG/L))

ND
D<=2
2<-5

5<=10
10<=20
>20
Total
Count
%
Count
%
Count
%
Count
%
Count
%
Count
%
Count
%
ND
79
122
17
026
16
025
12
0 19
17
026
12
0 19
153
236
D<=25
1554
2400
594
9!7
359
554
224
346
158
244
118
I 82
3007
4643
25<=120
614
948
481
743
400
6 18
213
329
187
289
190
293
2085
3220
120<=250
131
202
96
148
86
1 33
64
099
68
1 05
91
1 41
536
828
250<=500
115
1 78
68
105
41
063
28
043
25
039
32
049
309
477
>500
157
242
67
103
56
086
30
046
31
048
45
069
386
596
Total
2650
4092
1323
2043
958
1479
571
882
486
750
488
754
6476
1 00 00
Frequency Missing =1301
Note D = detect, ND = non-detect

Ground water station counts were compared to ground water observation counts to see if the
percentages changed significantly based on these two alternative counting techniques If a large
percentage difference is observed, then surface water station counts would also be made  Exhibit
4-4 shows, arsemc-sulfate co-occurrence frequency counts for observations, and stations,
respectively  A comparison of the differences in the totals column (for arsenic levels) indicates
that the difference between the two is small
        Exhibit 4-4. Comparison of Co-occurrence Frequency Percentage Based on
                               Observations and Stations
sulfate

Observations
Stations
ND
253
236
0-25
4346
4643
25-120
3336
3220
120-250
845
828
250-500
473
477
>500
720
596
May 21,1999
4-3
Draft Report

-------
                                                                              4  Results
The differences in the percentages are low, less than 0 5% in all cases by one, thus it appears that
co-occurrence frequency based on observations would yield similar conclusions as co-occurrence
frequency based on stations

4.2 Constituent Occurrence with Depth

The overall results of the well depth frequency analysis are documented m Appendix C Well
depth and constituent count was examined for all constituents at all threshold levels  Exhibit 4-5
summarizes well depth frequency percentage for the primary and  secondary constituents  In
general, constituent counts decrease with increase in well depth All constituents have high
counts between depth 100 feet and 250 feet

     Exhibit 4-5. Constituent Occurrence with Depth (percentage of observations)
Depth (feet)

Arsenic Diss
Arsenic Total
Sulfate Diss
Uraniurn-235 Diss
Uramum-238 Diss
Radium-226 Diss
Radmm-226 Diss
Radium-226PE
Radon-222 PE
Radon-222Totat
Antimony-Diss
Antimony Total
Barium Diss
Barium Susp
Barium Total
Beryllium Diss
Beryllum Total
Cadmium Diss
Cadmium Total
Chromium, Diss
Chromium Hexval
Chromium Total
Cyanide Total
Cyanide Diss
1000
410
92
5 10
1 85
1 85
2632
1121
1928
1 82
1 79
383
1637
532
0
720
380
16 19
296
978
348
609
886
1540
132
Sample
Size
8470
946
23107
54
54
76
107
166
4442
4518
2379
397
6731
1
667
4998
525
7778
910
8217
838
1084
474
1366
N.D.
3521
569
782
8
5
37
8
5
2
63
2310
373
495
0
291
4738
486
7019
848
539!
305
558
456
1338
May 21, 1999
4.4
Draft Report

-------
                                                                             4  Results
      Exhibit 4-5. (continued) Constituent Occurrence with Depth (percentage
                                    observations)
                                  of
Depth (feet)

Dissolved Solids
Hardness
Iron Susp
Iron Total
Iron Dtss
Iron Ferrous
Manganese Susp
Manganese Total
Manganese Diss
Mercury Diss
Mercury Total
Nickel Diss
Nickel Total
Nitrite
Nitrite
Nitrate
Thallium Diss
Thallium Total
Selenium Diss
Selenium Total
(K=50
2773
2778
0
22.25
2739
2877
00
25 16
2763
2847
2776
320
2670
3063
3496
4589
2003
3393
2685
3670
50<=100
1962
23 13
0
1354
2438
4041
8333
18 10
24 15
2757
1505
2344
1342
21 60
1289
2087
1362
2143
21 51
15 13
100<=250
2284
2249
100
1807
2360
1575
1667
2299
2382
22 41
1561
2340
1121
2286
1605
1586
2436
1643
2676
1438
250<=500
1374
1249
0
1885
1263
205
00
1936
1300
1323
2096
11 26
1770
1294
989
893
2260
393
1563
1337
500«=1000
1037
918
0
1699
838
1301
00
8 11
817
524
1616
5 17
1711
943
774
699
705
321
589
971
>1000
570
493
0
1030
363
00
00
629
324
308
446
473
1386
253
1848
1 47
1234
2107
336
1072
Sample
Size
18258
22067
1
1942
19744
146
6
1431
20024
2691
897
5069
678
20560
698
4692
624
280
6607
793
N.D.
12424
2
0
120
5768
13
0
364
5996
2650
867
3243
437
16487
590
224
565
278
5006
608
Note N D = Non Detects

4.3 Ancillary Parameters

Appendix D contains the results of a statistical summarization of eight ancillary parameters
(turbidity, conductance, dissolved oxygen, pH, hardness, alkalinity, well depth, depth below
land) for each EPA region For each parameter, minimum, maximum, mean, frequency, range,
and standard deviation were computed by EPA Region  The rationale for this analysis was to
serve as a first order indication of the variability in specific geochemical and hydrogeologic
conditions that might lead to a better understanding of the environmental conditions causing
certain parameters to co-occur
May 21, 1999
4-5
Draft Report

-------
                                                                           4  Results
4.4 Correlation Analysis

441  Correlation Coefficient Calculations

Correlation coefficients were calculated for all combinations of primary and secondary
constituents at all threshold levels Calculations were made separately for ground water and
surface water on both a national and regional basis  The detailed results of these computations
are shown in Appendices E (national level) and F (regional level)  Each appendix is divided into
three sets of results as follows

Appendix E - overall: national correlation coefficient analysis for ground water and surface
water irrespective of threshold level  Three values are reported for each constituent pair the
correlation coefficient (top), p-value (middle), and number of observations (bottom) Exhibit 4-6
shows a sample page from this appendix

Appendix E - threshold: national level results for correlation coefficients and summary
statistics by threshold level Exhibit 4-7 shows a sample page from this appendix  Each table m
this appendix contains the following information

VTAG = Constituent 1, short name
VTHRES = Constituent 1, threshold level
WTAG = Constituent 0, short name
MEDIUM = ground water or surface water
VARIABLE = Constituent  1, parameter number
VLABEL = Constituent 1, parameter name and units
WITH = Constituent 2, parameter number
WLABEL = Constituent 2, parameter name and units
VN = Constituent 1, number of observations
VMEAN = Constituent 1, mean
VSTD = Constituent 1, standard deviation
VSUM = Constituent 1, sum of values
VMIN = Constituent 1, minimum value
VMAX = Constituent 1, maximum value
WN, WMEAN, WSTD, WSUM, WMIN, WMAX = Similar statistics for Constituent 2
CORR = Correlation Coefficient
PVALUE = Measure of Significance
CORRN = Sample size used to calculate correlation
SIG = Indicator of Significance (* = significant, blank = not significant)

Appendix E - significant* national level results showing only significant correlation
coefficients  The significant correlations were extracted from Appendix E - threshold, and

May 21, 1999                              4^6                                Draft Report

-------
                                                                              4  Results
placed in these tables The headings m these tables, VTAG, VTHRES, VARIABLE, WTAG,
WITH, CORRN, CORK, and PVALUE are similar to those explained above  Exhibit 4-8 shows
a sample page from this appendix

These national level results serve as a first order indicator of co-occurrence potential  The data
are examined for the entire U S without regard for variation in environmental conditions that
may influence co-occurrence on a regional basis  In recognition of this, the data was partitioned
into 10 regions (see Exhibit 2-4) and the correlation analysis was repeated The results are
documented in Appendix F  The tables are similar in structure to Appendix E; one set per
region

442 Screening of Results

Calculation of the correlation coefficient and associated P-value was used as a screening
technique to filter out constituents which do not co-occur on a statistically significant basis  Any
pair of constituents with P-value greater than 0 05 was not considered for further analysis
Exhibit 4-9 shows all the constituents that pass through this filter on a national and regional
level In this exhibit, constituent pairs are listed irrespective of the threshold value at which they
occur The national level analysis acts as initial screening technique  It is first cut at identifying
non-random patterns of co-occurrence Regional analysis further refines the patterns at the
national level by identifying  specific geographic regions of co-occurrence The astensk (*) sign,
in this exhibit, indicates co-occurrence in that particular geographical unit A blank indicates no
pairs co-occur on a statistically significant basis in that region
May 21,1999                                4-7                                 Draft Report

-------
                                                                        4  Results
               Exhibit 4-6, Sample Page from Appendix E - overall. National Correlation Analysis Results
               *    i'                              Ground Water
•^
"* '- \'
i r
— V *
ALKALINITV-D FE-P39036-(MG/L AS
CAC03) j ,
ANTIMONY DISSOLVED P01095-
(UGA.ASS8) 1 -5
ANTIMONY TOTAL P01097-(UG/L AS
SB) , t
ARSENIC DISSOLVED-POIOOCHUGrt.
AS AS)
ARSENIC-SUSPENDED P01001
(UG/LASAS)
ARSENIC-TOTAL P01002HUG/L AS
AS)
BARIUM DISSOLVED P0100S-(UG/L
ASBA)
BARIUM SUSPENDED-P01008-(UGrt.
ASBA)
BARIUM TOTAL-P01007-(UG/L AS
BA)
B6RYIUUM-OISSOLVED-P01010
(UGA.ASBE)
BERYLLIUM TOTAL P01012-{UG/L
AS BE)
CADMIUM DISSOLVED-P01025
(UG/LASCD)
ALKAUNIT
"Y-OFE
rP39036
(MG/LAS
CAC03)
!:>Y
14--> •'
•*! rtf}-,^
o! i jrt "
0311
oooo
N-149

0133
0004
N=480
0106
0091
N=256

0085
0107
N=362
0472
0000
N=303
•0005
0919
NM04
0038
0544
N>=256
ARSENIC-
DISSOLVE
D-P01000-
(UGfl-AS
AS)
0341
0000
N=1312
0006
0741
N=2878
0311
0000
N=149
» l"\ ' '
^i*'J
'M' 4«?v


0009
0484
N=6087

0045
0598
N=142
0002
0894
N=4933
0331
oooo
N=145
OOOO
0979
N=7590
ARSENIC
SUSPEND
ED
P01001
(UG/IAS
AS)




.'Jf^f "
t **Vfife







ARSENIC
TOTAL
P01002-
(UGn.AS
AS)
0198
0431
N=ia
0132
0169
N=111
0133
0004
N=480


iff "m
m^
0042
0447
N=336
na
0023
OS42
N=715
•001S
0793
N<523
0118
0004
N=593
-0006
0915
N=329
BARIUM
DISSOLVE
D-P0100S-
(U6A.AS
BA)
0125
OOOO
N=1430
-0017
0365
N=2S28
0106
0091
N=2S6
-0009
0484
N=6087

0042
0447
N=336
*#
^ypj'fes


-0015
0235
N*6172
-0033
0598
N=265
0005
0687
N=6732
BARIUM-
SUSPEND
ED- '
P01006-
(UG/LAS
BA)t,





-na






BARIUM
TOTAL Y
P01007-
(UG/LAS
8A) >
-0117
0883
H=4
-0106
0292
N=101
0085
0107
N=362
-0045
0598
N=142

0023
0542
N=715


w&
•0032
0640
N=215
0037
0450
N=410
-0026
072t
N=191
BERYLLIU
- f M-
OISSOLVE
0-P01010-
(UGA.AS
, BE)
0 026 0 470
N=794
0166
DOOO
N=2935
0472
OOOO
N=303
0002
0694
N=4933

•0015
0793
N-323
-0015
0235
N=6172

-0032
0640
N=215
S^*f?'4!

0127
OOQO
N-6266
BERYLLIU
M-TOTAL-
P01012-4
(UQA.AS
•f "
f$»
-0165
0526
N=17
-0468
OOOO
N=98
-0005
0919
N-J04
•O331
OOOO
N=145

0118
0004
N=693
-0033
0598
N-265

0037
0450
N-410

1
1

I

m

0001
0989
N=258
CADMIUM-
DISSOLVE
D-P01025-
(UG/LAS
'**»- >
0015
0651
N=863
0006
0736
N-2939
-0038
0544
N-256
OOOO
0979
N=7590

-0006
0915
N=329
•0005
0687
N=6732

-0026
0721
N=191
0127
OOOO
N=6266
0001
0989
N-258
•
CADMIUM-
SUSPEND
ED--
TO102S-*
(UG/LAS
CO)












CADMIUM-
TOTAL-
P01027-
(UGrt.A3
CD)
0577
0015
N=17
0328
0002
N=87
0461
OOOO
N-438
0239
OOOO
N=263

-0030
0368
N=899
-0014
0776
N=401
na--
0048
0199
N=705
0285
OOOO
N=388
0030
0404
N=S51

CHROMIU
M
DISSOLVE
D-P01030-
(U&LAS
* CR)
0033
0340
N=B51
0330
OOOO
N-2920
0935
OOOO
N=266
0043
OOOO
N^7641

0284
OOOO
N=389
-0022
0070
N=6503
na
0055
0441
N=201
0767
OOOO
N=5981
0029
0640
N=266
0090
OOOO
N*8965
May 21, 1999
4-8
Draft Report

-------
                                                                          4  Results
    Exhibit 4-7. Sample Page from Appendix E - threshold. National Correlation Analysis.  Correlation Coefficients and
                                  summary statistics by threshold level (ground water)
VTAQ
ALKALINITY
ALKALINITY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
1/THRES
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
a
a
0
D
0
0
0
0
0
0
0
WTAG
CONDUCTANCE
TURBIDITY
ALKALINITY
ALKALINITY
BARIUM
BARIUM
BARIUM
BARIUM
BARIUM
BARIUM
BERYLLIUM
BERYLLIUM
BERYLLIUM
BERYLLIUM
CADMIUM
CADMIUM
CADMIUM
CADMIUM
CHROMIUM
CHROMIUM
CHROMIUM
CHROMIUM
CHROMIUM
CHROMIUM
CONDUCTANCE
CONDUC TANC E
MEDIUM
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
GROUND WATER
VARIABLE
P39036
P39036
P01095
POI 097
PQ1095
P0109S
PQ1095
POI 097
PQ1097
P0tOfl7
P0109S
P01Q95
POI 007
P01097
P01095
P01095
POI 097
P01M7
POI 095
P0109S
P01 095
PQ1Q87
P01097
POI 097
POI 095
PQ1097
VLABEL «
ALKALINITY o FE P39036
(MG/U AS CACO3)
ALKALINITY O FE P39036
(MG/L AS CACO3)
ANTIMONY DISSOLVED
POtQBS [UGfL AS SB)
ANTIMONY TOTAL P01097
(UG/L AS set
ANTIMONY DISSOLVED
P01095(UC/IASSB)
ANTIMONY DISSOLVED
P01095-(UG/LASSB>
ANTIMONY DISSOLVED
P01Q95(UG/LASSB)
ANTIMONY TOTAL P01097
(UG/L AS SB)
ANTIMONY TOTAL P01097
(UG/L AS SB)
ANTIMONY TOTAL PQ1097
(UG/L AS SB)
ANTIMONY DISSOLVED
POtOSS (UG/L AS SB)
ANTIMONY DISSOLVED
P01095-(UG/LASSB)
ANTIMONY TOTAL P01097
(UG/t AS SB)
ANTIMONY TOTAL potos?
(UG/L AS SB)
ANTIMONY DISSOLVED
PQ109^(UG/L AS SB)
ANTIMONY DISSOLVED
P01085-JUG/IASSB)
ANTIMONY TOTAL P0109?
(UG/L AS SB)
ANTIMONY TOTAL P01097
(UG/L AS SB)
ANTIMONY DISSOLVED
P01 095 (UG/L AS SB)
ANTIMONY DISSOLVED
P01095(UG/1_ASSB)
ANTIMONY DISSOLVED
P01OT5{UGAASSB)
ANTIMONY TOTAL P01097
(UC/L AS SB)
ANTIMONY TOTAL P0l097
(UG/L AS S8)
ANTIMONY TOTAL P0109T
(UG/L AS SB)
P01095 IUG/L AS SB)
ANTIMONY TOTAL P01097
(UG/L AS SB)
WITH
P00095
P00076
P39B36
P38036
PQ1QQ5
P01006
PQ1007
P01005
PQ1D06
P01007
POtOlO
PQ1012
P01010
PQ1Q12
P01025
P01027
PD1Q25
PQ1027
P01030
P01032
P01034
P01Q30
PC1D32
P04034
P00095
P00095
WLABEl -?^^-iv -^ ^^j^v 3
SPECIFIC CONDUCT POG095 (US/CM Q 25C)
TURBIDITY PQ007G (NTUJ
ALKALINITY D FE P39036 {MG^, AS CACQ3)
ALKALINITY D FE P390M 
BARIUM SUSPENDED POI006 (UG/L AS BA)
BARIUM TOTAL P01007 (UGfl. AS BA)
BARIUM DISSOLVED-P01005 (UG/L AS BA)
BARIUM SUSPENDED-P01 006 (UG/L AS BA)
BARIUM TOTAL PQ1QQ7 (UGiASBAJ
BERYLLIUM DISSOLVED POHJ10-{U&T. AS BE)
BERYLLIUM TOTAL P010f 2 (UGA. AS BE)
BERYLLIUM DISSOLVED P010)0-{UG/L AS BE)
BERYLLIUM TOTAL P01Dt2 (UG/L A9 BE)
CADMIUM DISSOLVED-P01D2S-(UG/L AS CD)
CADMIUM TOTAL PQ1027 (UGA. AS CD)
CADMIUM DISSOLVED POI 025 (ucn. AS CD>
CADMIUM TOTAL P01027 (UG/L AS CD)
CHROMIUM DfSSOLVED-P01030-(U&0. AS CR}
CHROMIUM HEXAVAL PoiD32 (UG/L AS CRJ
CHROMIUM TOTAL POI 014 (UG/L AS CR)
CHROMIUM DISSOLVED PQ103Q (UG/L AS CR)
CHROMIUM HEXAVAL P0103Z (UG/L AS CR)
CHROMIUM TOTAL PQ1Q34 (UG/L AS CR)
SPECIFIC CONDUCT P00095-(US/CM Q 25C>

VH j, r
2642
2642
2940
4B2
2MO
2940
2940
492
402
482
2940
2940
462
4H2
2940
2940
482
4&2
2940
2940
2940
402
4B?
482
2940
482
VMCAN ;i
204 9232
204 9232
7 838537
2S 07158
7 838537
7 039537
7 838537
28 07158
2607158
2& 07 15ft
763654
763854
28 071 $6
2807T58
7 638537
7 83BS37
28 071 58
28 071 $6
7 638537
7 6*8537
7 838537
28 07 i $8
2807156
2807158
7BMS37
2807158
VSTD f
146 6665
146866S
2993498
61 3556
29 934 96
29 93498
29 93498
01 3S5B
61 3556
61 3556
29 9)498
29 93498
6t 3556
61 3558
29934M
29 93496
61 3558
6t 3556
2993498
29»4M
29 93498
61 3556
61 3SS6
61 35SA
2993498
6t 3SS6
VSUM
S41407
$41407
2304S
13531
23045
23045
23045
t3S31
13S31
1353t
23045
23045
13531
13531
23045
23045
13531
13531
23045
23045
23045
13531
13531
13531
23045
13531
WIN
3
3
1
1
1
1
1
t
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
VMAX
4500
4500
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
WN
56711
2359
2642
2942
8200
2
776
8200
2
776
6332
627
6332
627
9321
1131
9321
1131
9790
052
127«
9790
952
1274
S671*
567(1
WMEAN
1443 2B7
2336229
204 9232
204 9232
120 7616
100
194384
1207646
100
194 384
1 869
352998
1 869
3 52998
7 570965
4 464633
7570965
4464633
9 877542
1300315
135
9 877542
13 00315
13 5
1443287
1443267
May 21, 1999
4-9
Draft Report

-------
                                                                           4  Results
  Exhibit 4-8.  Sample Page from Appendix E - significant.  National correlation analysis.
              Significant correlation coefficients for ground water samples
VTAG
ALKALINITY .
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
ANTIMONY
VTHRES
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
VARIABLE
P39036
P01095
P01095
P01097
P01095
P01097
P01095
P01095
P01097
P01097
P01095
P01095
P01097
P01095
P01095
P01097
P01095
P01095
P01097
P01095
POI095
P01095
P01097
P01095
POI095
P01097
P01097
POI095
POI095
P01097
P01095
P01097
POI095
P01095
WTAG
CONDUCTANCE
BERYLLIUM
BERYLLIUM
BERYLLIUM
CADMIUM
CADMIUM
CHROMIUM
CHROMIUM
CHROMIUM
CHROMIUM
CONDUCTANCE
CYANIDE
CYANIDE
DEPTH-BELOW-LAN
DISSOLVED-OXYGE
DISSOLVED-OXYGE
DISSOLVED-SOL1D
HARDNESS
HARDNESS
IRON
IRON
IRON
IRON
MANGANESE
MERCURY
MERCURY
MERCURY
NICKEL
NICKEL
NICKEL
NITRITE
NITRITE
SELENIUM
SELENIUM
WITH
P00095
P01010
POI012
P01010
POI027
P01027
P01030
P01034
P01030
P01034
P00095
P00720
P00720
P72019
P00300
P00300
P70301
P00900
P00900
P01045
P01046
P01047
P01045
P01056
P71890
P71890
P71900
P01065
P01067
P01067
P00615
P006I5
P01145
P01147
CORRN
2594
2935
98
303
87
438
2920
114
266
461
2736
53
178
1029
2369
189
2360
2689
275
292
2698
75
449
2871
796
73
363
2827
116
461
83
198
2727
101
CORR
0 13978
0 I658I
-0468
047221
032791
0461
032964
03737
0 93534
0 33479
0 09728
-068331
0 99895
024349
-009713
0 32406
0 26347
0 44634
-0 17851
021065
0 65222
0 34914
0 18448
0 1 7007
0 10665
037241
028514
036643
029066
023447
-0 94846
0 24608
018984
0 94559
RVALUE
00001
00001
00001
00001
00019
0000 1
00001
00001
00001
00001
00001
0000 1
00001
00001
00001
00001
00001
00001
0003
00003
00001
00021
00001
00001
00026
00012
00001
0000 1
00015
00001
00001
00005
00001
00001
May 21, 1999
4-10
Draft Report

-------
                                                                               4  Results
                                  Exhibit 4-9. Statistically Significant Co-occurring Pairs


Antimony


















Arsenic









Arsenic
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nicket
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-235
Uramum-238
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Ground Water
National
*

*
*
*
*
*
*
*
*
4
*


*
*
*
*
*
*

*
*
*

*

1

























*

2

*
*
























3





















*

*



4
*
*



*


*

*



*
*



*
*


*
*
*

5

*



*


*





*
*




*




*

6
*










*



*



*







7
*
*



*




*
*


*


*
*
*




*
*

8
*





*



*

*


*

*
*
*







9
*










*


*
*

*
*
*







10

*



















*





Surface Water ,
National
*
*


*
*
*
*
*

*



*
*
*


*

*
*

*
*
*
1
*









*




*



*







2
*
*
*
*
*

*
*

*









*

*
*
i"1

*
*
3














*







*
*


*
4
*




*
*
*











*

*
*

*
*
*
5






*
*


*
*




*








*
*
6
















*




*
*




7





*
*
*


*
*




*




*
*


*
*
g






4
*


*
*




*




*
*

*
*
*
9






*
*


*
*








*
*
#


*
*
10






*
*



















May 21, 1999
4-11
Draft Report

-------
                                                                               4  Results
                                 Exhibit 4-9. Statistically Significant Co-occurring Pairs












Barium


















Mercury
Nickel
Nitrate
Nitrtte
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-235
Arsenic
Antimony
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Ground Water
National
*
*
*

*
*
*

*





*
*
*
*
*



*
*
*
*

1







*




*







*




*

2

*

*







*
*
*
*



*


*
*
*



3

*




*





*
*
*







*

*
*

4

*

*


*
+
*

*
*
*
*
*
*


*


*


*


5

*


*

*
*


*
*






*

*

*


*

6




*

*

*



*











*


7

*

0




*


*
*
*
*
*


*
*




*
*

8






*
*





*













9

*

*



*






*


*

*







10


*
*


*




*

*
*


*


*



*


Surface Water
National

*
*


*
*
*
*


*
*
*
*
*
*
*
*
*
*
*

*
*
*
*
1


*




*





*



*

*
*
*



*

2







*



*

*


*
*

*
*


*



3













*






*






4
*

*





*









*








5
*



















*






6
*

*


*


*









*

*






7
*

*





*



*





*

*





*
8






*

*



*


*


*

*


*



9
*

*


*
*





*





*





*

*
10

*

*

*






*







*
*
*




May 21, 1999
4-12
Draft Report

-------
                                                                               4  Results
                                  Exhibit 4-9. Statistically Significant Co-occurring Pairs




Beryllium


















Cadmium







Uramum-235
Uramum-238
Arsenic
Antimony
Barium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-235
Uranmm-238
Arsenic
Antimony
Barium
Beryllium
Chromium
Cyanide
Ground Water
National


*
*

*
*
*
*
*
*
*

*
*

*
*
*
*

*
*

*
*
*
1




*

*


*

*
*














2



*
*
*







*









*
*


3


*

*




*




*


*





*



4




*
*

*

*


*



*






*
*

*
5





*



*






*







*


6

*


*
*



*


*
*






*



*


7




*


*




*
*





*
*


*


*
8





*



*


*
*
*




*
*


*
*


9





*
*
*











*
*



*

*
10


*


*
*

*
*

*





*

*



*
*


Surface Water
National


*

*
*
*

*
*
*
*
*
*



*
*


*

*
*


1




*
*






*










*
*


2


*
*
*
*










*
*



*
*
*
*


3




*
*











*



*

*
*


4


*




*





*


*

*


*





5

















*









6


*










*


*
*



*



+

7


*










*


*

*


*





8

*
*










*


*
*
*


*



*
* \
9


*

*


*






*

*
*



*



*

10




*
*


*
















*

May 21, 1999
4-13
Draft Report

-------
                                                                              4   Results
                                 Exhibit 4-9. Statistically Significant Co-occurring Pairs















Chromium















Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Ground Water
National
*
*
*
*

*
*

*
*
*
*
*
*
*
*
*
*
*
*

*
*

*
*

1
*















*










2
*

*

*
*









*

*

*




*


3
*







*






*

*







*

4
*



*



*






*


*








5








*


















6























*

*

7




*
*





*
*


*


*




*



8




*
*
*




*
*



-










9






*




*
*


*
*

*






*

10



*


*
*



*



*
*


*
*




*

Surface Water
National


*
*
*
*




*


*
*
*
*

*
*
*
*

*
*


1



















*
*



*


2
*




*



*




*




*







3









*













*



4


*


*



*














*


5









*

















6


*


*




*












*



7


*


*




*





*




*

*



8

*
*


*




*





*
*
*


*




*
9

*
*













*
*



*

*


*
10
*
4





*





*



*







*
*
May 21, 1999
4-14
Draft Report

-------
                                                                             4  Results
                                 Exhibit 4-9. Statistically Significant Co-occurring Pairs







Cyanide


















Iron




Selenium
Sulfate
Thallium
Urarnum-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uranmm-235
Uramum-238
Arsenic
Antimony
Barium
Ground Water
National
*

*
*
*

*
*
*
*
*
*
*
*
*
*
*


*
*
*


*
*
*
1
























*


2
*


























3



























4





*
*
*
*
*
*
*
*
*
*




*

*


*


5






*







*




*
*



*


6




*






















7



*
*
*
*
*
*
*
*





*


*
*
*


*


8

























*

9

*

*
*



*
*
*
*
*
*
*
*
*


*
*






10

*

*















*
*




_v

Surface Water
National

*
*
*
*
*
*
*


*


*


*


*




*
*
*
I



























2
























*
*
*
3



























4


*


+
+

*















*
*

5












*
*
*




*




*
*

6



























7


*



*






*






*



*
*

8

*
*


*

*

*
*
*



*
*


*
*



*
*

9
*
*

*
*



*















*
*

10




%




















*

May 21, 1999
4-15
Draft Report

-------
                                                                               4  Results
                                 Exhibit 4-9. Statistically Significant Co-occurring Pairs


















Manganese












Beryllium
Cadmium
Chromium
Cyanide
Manganese
Mercury
Nickel
Nrtrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Mercury
Nickel
Nrtrate
Ground Water
National
*
*
*
*
*
*
*


*

*
*
*



*
*
*
*
*
*
*
*
*
*
!



















*







2







*




*














3







*

*


*






*







4



*















*

*
*

*


5









*









*




*


6







*


*


*





*




*

*
7








*
*











*





8








*

*


*





*





*

9



*







+



*


*

*
*
*

*
*

10
*

*



*


*








*
*
*




*
*
Surface Water ,
National
*

*


*
*

*




*


*
*
*
*
*
*

*

*

1


*




*
*









*







*
2

*
*




*


*
*
*



*
*
*







*
3







*



*
*



*









*
4





*



*


*



*
*









5







*



*
*



*
*




*



*
6





*


















*


7





*
*
*



*
*
%


*
*








*
8
*


*

*
*
*




*



*
*

*
*
*




*
9
*




*
*
*



*
*



*
*


*





*
10
*





*

*



*




*


*


»



May 21, 1999
4-16
Draft Report

-------
                                                                               4   Results
                                  Exhibit 4-9. Statistically Significant Co-occurring Pairs










Mercury




















Nitrite
Radon
Radium
Selenium
Su IFate
Thallium
Uranium-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uranium-235
Uramum-238
Ground Water
National

*
*
*
*
*
*


*
*
*
*
*
*
*
*
*
*
*


*

*


1




*






















2




*





*

*














3

*
*

*












*









4









*
*

*

*

*
*

*


*

*


5




*




*
*

*



*
*




*




6




*







*



*







*


7
-
*








*

*




*









8

*
*

*
*






*




*
*
*







9
*

*
*
*
*

*






*

*
*
*
*
*

*
*



10

*
*

*

*












*







Surface Water
National




*
*


*
*
*
*
*
*
*
*

*
*



*
*
*


1
*


























2
*


*




*


















3
*


*




*


















4
*
*






*

+

*


*

*




*




5



*




*





*







*




6










*

*


*
*
*
*
*


*
*
*


7





*


*

*

*
*
*
*

*





*



8


*

*



*

*

*
*

*

*
*



*
*
*


9



*
*



*

*

*
*

*











10

*
*

*


*



















May 21, 1999
4-17
Draft Report

-------
                                                                               4  Results
                                  Exhibit 4-9  Statistically Significant Co-occurring Pairs


Nickel


















Nitrate









Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uranmm-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Ground Water
National
*
*

*
*
*
*
*
*
*


*

*
*
*
*


*




*

1



*

















*
*




2
*










*



*







*


*
3
*








*





*










*
4
*





*


*


*







*

*
*



5
*





*


*






*




*





6


















*



*

*

*
7
*

4






*







*
*

*

*
*
*


8








*
*


*




*


*

*
*



9
*

*



*

*
*

*
*



*

*






*

10



*
*


*
»





*
*

*



id





Surface Water
National
*

*
*
+


*
*
*



*
*
*
*
*
*

*
*
*
*
4


1


*

*



*


*



*




*
*
*



*
2

*
*

*



*






*





*




*
3




*










*





%


*

*
4




*




*

*















5






*




*








*
*




*
6









*

*

*







*


*


7









*






*



4
*


*

*
8









*

*

*
*





*
*



*
*
9

*









*


*


*


^



*

*
10

*









*









*





May 21,1999
4-18
Draft Report

-------
                                                                              4  Results
                                 Exhibit 4-9. Statistically Significant Co-occurring Pairs













Nitrite

















Manganese
Mercury
Nickel
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uranium-235
Uranmm-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Radon
Radium
Selenium
Sulfate
Ground Water
National
*
*
*
*


*
*
*
*
*

*

*
*
*
*


*

*
*

*

1







*



















2






*




*

*
*
*
*




*





3



*


















*
*



4




*

*
*



*

it






*






5

























*

6
*









*

*

*












7


*








*
*

*
*

*
*




*

*

8

*

*



*






*
*


*

*

*
*

*

9

*

*



*



*
*




*

*
*
*
*
*

*

10
*


*



*



*








*

*


*

Surface Water
National

>t
*
*
*


*
*




*
*
*
*
*
*



*




1
*

*




*





*


*

*
*

*


*


2
*

h*


*

*







*



*







3
*

*
















*







4






*
*
*





*
*
*


*

*



*

5
*




*
*
*




*








*



*

6

*
*



*







*
*




*
*





7
*



*



*



*

*
*











8
*
*
*



*
*
*



*

*
*

*



*


*
*

9
*

*



*
*




*








*


*
*

10







*



*

*




*


*



*

May 21, 1999
4-19
Draft Report

-------
                                                                               4  Results
                                  Exhibit 4-9. Statistically Significant Co-occurring Pairs





Radon


















Radium






Thallium
Uramum-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radium
Selenium
Sulfate
Thallium
Uramum-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Ground Water
National
*


*

*
*
*
*

*
*

*

*


*



*

*


i



























2





*


















*


3





*
*

*

*
*



*


*








4













*
a












5



*

*




*
















6



*




*


















7










*
*



*

*









8




*

+
+



*

*

*


*








9







*
*



*
*

*

*
*

*
*





10







*
*

*
*






*







*
Surface Water
National
*













*


*
*



*

*


1



























2
























*


3



























4
*









*
*






*








5



























6
*





















*




7
*













*



*








8
*























*


9
*





*










+
*



*




10





*


*


*










*



*
May 21,1999
4-20
Draft Report

-------
                                                                               4  Results
                                  Exhibit 4-9. Statistically Significant Co-occurring Pairs
















Sulfate







_,






Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Selenium
Sulfate
Thallium
Uranmm-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Ground Water
National


-
*
*




*
*
*

*

*
*
*
*

*
*
*

*
*
*
1
















*





*


*

2




*
















*
*




3



*












*
*



*
*



*
4















*









*

5















*
*



*
*
*




6


*












*





*
*




7
















*



*
*





8


*
*











*





*
*


*

9



*





*

*



*



*
*
*
*
*

*

10



*








*
*





*
*
*
*


*

Surface Water ,
National





*



*





*
*


*

*
*
*

*

1







*







*
*







*
*

2


*



*



*




*

*



*


*
*

3

















*



*


*


4





















*



*

5






*








*

*



*



*

6





*




*




*

*





*



7




















*
*

*


*
8
*


*

*

*


*




*

*

*
*
*
*
4

*
*
9
*






*







*

*
*
*

*
*


*
*
10
*


*









*




*


*
*


*
*
May 21, 1999
4-21
Draft Report

-------
                                                                              4  Results
                                 Exhibit 4-9.  Statistically Significant Co-occurring Pairs








Thallium


















Uramum-235



Radon
Radium
Selenium
Thallium
Uramum-235
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Uramum-235
Uramum-238
Arsenic
Antimony
Ground Water
National
*
*
*

*
*
*
*

*
*
*
*
*
*
*
*
*
*

*



*

*
1



























2



























3
*


























4






*





*


*











5
















*










6



*


*






*

*






*




7



*
*
*
*





*









*



*
8
*


*
*
*







*
*







*



*
9
*

*
*

*








*

*



*

*



*
10
*

*

*






















Surface Water
National
*

*
*


*
*
*
*
*
*

*
*
*
*
*
*



*




1



























2

*
*
























3



























4
*


*


*


*

*





*
*



*




5







*



















6

*




4
*


*




*


*








7
*


*


*
4
*
*
*
*

*
*

*
*
*



*




8

*

*

*
*
*

*
*
*



*

*
*



*




9
*







*









*








10



























May 21, 1999
4-22
Draft Report

-------
                                                                              4  Results
                                 Exhibit 4-9.  Statistically Significant Co-occurring Pairs



















Uramum-238











Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-238
Arsenic
Antimony
Barium
Beryllium
Cadmium
Chromium
Cyanide
Iron
Manganese
Mercury
Ground Water
National

*
*
*


*

*

*


*


*

*


*
*




1



























2



























3



























4



























5



























6













*





*
*

*




7

*
*
*




*




*




*

*
*
*




8

*
*





*









*

*
*





9

*
*
*














*

*
*
*

*
*

10

*
if
*


*

*


















Surface Water
National



*




+













*




1



























2



























3



























4



























5



























6
















*










7



























8



















*







9



*




*







*





*




10






















*


*

May 21, 1999
4-23
Draft Report

-------
                                                                                4  Results
                                  Exhibit 4-9.  Statistically Significant Co-occurring Pairs













Nickel
Nitrate
Nitrite
Radon
Radium
Selenium
Sulfate
Thallium
Uramum-235
Ground Water
National


*


*

*

1









2









3









4









5









6
*




*



7
*




*



8









9
*




*



10









Surface Water
National
*








1









2









3









4









5









6









7









8









9









10









May 21, 1999
4-24
Draft Report

-------
                                                          5  Summary and Conclusions
5. SUMMARY AND CONCLUSIONS

The purpose of this analysis was to determine whether specific drinking water contaminants
(primary and secondary) co-occur on a statistically significant basis or is the co-occurrence
purely a random phenomenon All the data for analysis was extracted from USGS NWIS (surface
and ground water) database Three main analyses were performed,

    •   Statistically significant correlation (national and regional level)
    •   Single and co-occurrence counts
    •   Constituent co-occurrence with well depth

For this project we analyzed a total of 22 constituents (7 primary and 15 secondary)  Several
constituents had  both total and dissolved forms  In addition, each constituent was analyzed at
one or more threshold levels   Based on these numbers, a total of 9,912 co-occurrence
combinations were analyzed on a national level  The screening techniques applied to the
correlation coefficients resulted in 1860 statistically significant pairs (based on P-value < 0 05)
The screening results for the regional analysis showed the following statistics 98,185
combinations and 7,152 statistically significant pairs based on p-value

The results from the statistical analysis shows the following conclusions,

    •   Statistically significant correlations were identified at the national and regional level
       Regional analysis further refined the patterns at the national level by identifying specific
       geographic regions of co-occurrence  Exhibit 4-9 presents an overall summarization of
       the co-occurrence patterns Examination of Exhibit 4-9 shows, for example, that in
       ground water, significant arsemc-sulfate co-occurrence appears in EPA regions 4,5, 8,
       and 9 For surface water, the arsemc-sulfate co-occurrence appears in regions 1,2,5,6,
       7, 8, 9, and 10 These conclusions are listed for all the combinations in Exhibit 4-9
       Exhibit 5-1 indicates the following patterns present at the national level and also in all the
       US EPA Regions in ground water and surface water
May 21, 1999                                 5-1                                 Draft Report

-------
                                                         5  Summary and Conclusions
     Exhibit 5-1. Constituent pairs which co-occur on a statistically significant basis in all
                                     10 EPA Regions
Ground- Water
Cadmium - Iron
Chromium - Nickel
Iron - Manganese





Surface Water
Arsenic - Nickel
Barium - Beryllium
Barium - Chromium
Chromium - Manganese
Chromium - Nickel
Iron - Manganese
Manganese - Nickel
Nitrate - Nitrite
       The results showed that the following pairs co-occur with radon in different EPA Regions
       in surface water

       Region 4 Radon - Iron, Radon - Manganese, Radon - Sulfate
       Region 7 Radon - Nitrate, Radon - Sulfate
       Region 9 Radon - Beryllium, Radon-Selenium, and Radon - Sulfate
       Region 10 Radon - Barium, Radon- Chromium, Radon - Manganese

       A comparison between the "observation count" tables with the "station count" tables
       (Exhibit 4-3) shows insignificant differences  Exhibit 4-3 shows examples of arsenic-
       sulfate co-occurrence frequency counts for observations, and stations, respectively

       As shown by the analysis, the number of constituent counts in general decrease with
       increase in well depth  This is not necessarily due to the fact that the constituents do not
       occur with depth  As shown in Exhibit 5-2, that there are fewer stations that tap deeper
       water Chromium hexval has  highest counts at the depth of 500-1000 feet
                Exhibit 5-2. Distribution of Stations by Well Depth (feet)
DEPTH
Number
Percentage
0-50
6280
26%
50-100
4616
19%
100-250
6489
27%
250-500
3293
14%
500-1000
2237
95%
>1000
931
4%
May 21, 1999
5-2
Draft Report

-------
                                                                      6 References
6. REFERENCES

Appelo, C A J and D Postma, Geochemistry, ground water, and pollution, A A Balkema,
1994

Deutsch, W  J ,Ground water geochemistry. Fundamentals and application to contamination,
Lewis Publishers, 1997

Dojhdo, J, and G A Best, Chemistry of water and water pollution, Ellis, Horwood, 1993

Fetter, C W., Contaminant Hydrogeology, Macmillan Publishing Company, 1993

Hoel, P G, Elementary Statistics, Wiley Senes in Probability and Mathematical Statistics, 1971

Murdoch, J and J A  Barnes, Statistical Tables for Science, Engineering, Management and
Business Studies, John Wiley and Sons, New York, 1974

SAIC, Co-occurrence of Drinking Water Contaminants, Final Draft Report, January 27, 1999

Toft, P, et al, Asbestos in Drinking Water, Cnt Revi  Environ Control, 14(2), 151, 1984

US EPA, National primary drinking water regulations  Contaminant specific fact sheets
Inorganic Chemicals - Technical Version EPA-811-F-95-002-T, 1995

Zuane, J D, Handbook of drinking water quality Standards and Controls, Van Nostrand
Remhold, New York
May 21,4 999                               6-1                                Draft Report

-------