|
NATURALLY OCCURRING CONCENTRATIONS OF INORGANIC CHEMICALS
<br /> IN GROUND WATER AND SOIL AT CALIFORNIA AIR FORCE INSTALLATIONS
<br /> Philip M.Hunter,P.G.,Air Force Center for Environmental Excellence,Brooks AFB,Texas 78235
<br /> Brian Davis,Ph.D.,Department of Toxic Substances Control,Sacramento,California
<br /> ABSTRACT
<br /> Risk assessment and risk management must differentiate between naturally occurring and anthropogenic inorganic chemicals.
<br /> Naturally occurring or background concentrations of inorganic chemicals are important for site characterization, determining
<br /> chemicals of concern,establishing cleanup levels,and long-term monitoring programs. Analysis of the Air Force's Environmental
<br /> Resources Program Information Management System(ERPIMS)database identified uncontaminated sample locations for soil and
<br /> ground water from 12 Air Force installations across 10 California counties. Background data for 25 inorganic constituents were taken
<br /> from 3000 borehole and 750 monitoring well locations. Maximum sample sizes for individual chemicals range from 1800 and 7400,
<br /> depending on the sampling medium.The 95'h percentile was used as the best statistic to represent background. Medians and 991h
<br /> percentiles of background levels are also presented. Since statistical analysis of soil data indicated that background levels differed
<br /> significantly with depth,separate background calculations for soil are presented for three depths(less than 2.5 feet,between 2.5 and
<br /> 10 feet,and greater than 10 feet). For ground water,background statistics for each constituent are given without regard to sampling
<br /> depth. Some inorganic constituents were detected frequently and at levels that exceed important environmental thresholds such as
<br /> Maximum Contaminant Levels (MCLS)or Action Levels for drinking water. Background 95'h percentile levels equal or exceed
<br /> federal and/or California MCLS for aluminum,antimony,arsenic,beryllium,cadmium,chromium,nickel,and thallium. The 95'h
<br /> percentile level for lead exceeds the U.S. EPA Action Level of 0.015 mg/L for drinking water measured at the tap. This analysis
<br /> provides background levels that are representative of California installations as a group. These data should not replace local
<br /> background data,but rather provide important benchmarks by which the adequacy of local data can be judged.
<br /> INTRODUCTION
<br /> Most contamination at Air Force Bases(AFBs)is organic,typically associated with chlorinated solvents and fuels. The presence of
<br /> key organic chemicals detected in ground water and soil samples is a good indicator of both inorganic and organic contamination.
<br /> Even when investigations specifically target contamination by drilling wells and borings into areas of hazardous waste sites,the non-
<br /> detect(ND)rates for organic chemicals are surprisingly high. The ND rates for trichloroethene,which is highly mobile and the most
<br /> ubiquitous constituent at AFBs,are about 65%in ground water. Other organic constituents have NO rates around 90%for ground
<br /> water. ND rates for organic chemicals in soil tend to be even higher. As a result and given the Air Force's extensive monitoring
<br /> network, an abundance of existing sampling locations are known to be uncontaminated and can be used to estimate background
<br /> concentrations. Computer algorithms were applied to the Air Force's Environmental Resources Program Information Management
<br /> System(ERPIMS)to identify background locations. Over 10 years of project data are available for background determinations at
<br /> California AFBs. This poster presents the automated approach to identify background locations, the statistics used to calculate
<br /> background concentrations,and background concentrations for both ground water and soil.
<br /> METHODS
<br /> A computer algorithm was constructed to identify background locations across all California AFBs. The algorithm,using Structured
<br /> Query Language (SQL), searches out all locations that have been sampled for both inorganic and organic chemicals. Sampling
<br /> locations with organic contamination are eliminated from the search.High-end outliers,which may represent locations contaminated
<br /> with inorganic chemicals but not organic chemicals,were also eliminated for each constituent based on"box and whisker"plots.Both
<br /> upgradient,downgradient,and sidegradient locations could potentially be identified as background sampling locations. Substantially
<br /> more background locations were identified in soil compared to ground water. On average,at least 25 background well locations and
<br /> 50 background borehole locations per AFB have been identified using these procedures. A large number of distinct sample locations
<br /> and large sample sizes easily meet the requirements for the statistical calculations used to determine background levels.
<br /> This analysis is complicated by multiple detection limits, diverse hydrogeologic terrains, variability over 3-dimensional space, a
<br /> variety of types of hazardous waste sites,multiple Air Force bases,and different waste handling practices. All of these issues force
<br /> one ultimately to discriminate background levels across more than one hydrostratigraphic unit or more than one soil horizon. The 95'h
<br /> percentile was used to best represent background for each analyte for ground water and soil. For small data sets,confidence limits are
<br /> important but for larger data sets as we have here, the simple percentile of the data is sufficient. In addition, the median (501h
<br /> percentile) and 991h percentile as well as sample size, number of well locations, number of AFBs, and detection frequency are
<br /> presented for each chemical. The number of wells and sample sizes of individual AFBs are presented. All statistical analysis used
<br /> SAS®and Systat®software.
<br /> BACKGROUND LEVELS FOR GROUND WATER
<br /> The analysis for California ground water is based on 2936 monitoring wells,sampled for both inorganic and organic chemicals,and
<br /> over 34,000 analytical records. The number of background wells analyzed varied from 94(chromium-6)to 653(lead);sample sizes
<br /> varied from 134(chromium-6)to 1819(magnesium);number of AFBs varied from 5(boron)to 12 for many chemicals. Background
<br /> level calculations are biased,with more than 50%of data from Vandenberg,Travis,and March AFBs. Detection rates in ground
<br /> water vary,with chloride,magnesium and sodium detected in 99%of samples,compared to 3%for cyanide and silver. Antimony,
<br /> arsenic,beryllium,cadmium,chromium-6,cobalt,copper,cyanide,lead,mercury,nickel,selenium,silver,and thallium had median
<br /> (501h percentile)concentrations less than the median method detection limits(MDLs). The 95'h percentile for cyanide and mercury
<br /> were also below MDL, indicating that they are rarely detected in ground water. Conversely, some inorganic constituents were
<br /> detected frequently and at levels that exceeded important environmental thresholds such as Maximum Contaminant Levels(MCLS)or
<br /> Action Levels for drinking water. The 95'h percentiles for arsenic, antimony,beryllium, chromium, nickel, and thallium exceed
<br /> MCLS. The 95'h percentile for arsenic is 0.05 mg/L,which is the current MCL. The background level for lead exceeds the U.S.EPA
<br /> Action Level of 0.015 mg/L for drinking water at the tap. Based on limited analysis,it can be said that chromium-6 is infrequently
<br /> detected in ground water.
<br />
|