AQUATIC EFFECTS RESEARCH PROGRAM
OVERVIEW
National Surface Water Survey
Eastern Lake Survey - Phase I
Synoptic Chemistry
Status and Extent
-------
DON No. 86-258-012-12-13
Work Assignment No. 12
EPA Contract No. 68-01-7288
AN OVERVIEW OF THE
NATIONAL SURFACE MATER SURVEY
EASTERN LAKE SURVEY - PHASE I
SYNOPTIC CHEMISTRY
A Contribution to the National Acid Precipitation
Assessment Program
Prepared for:
Dr. Rick A. Linthurst
Director, Aquatic Effects Research Program
U. S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
Research Triangle Park, North Carolina 27711
Prepared by:
Penelope Kellar
Radian Corporation
3200 Progress Center
Research Triangle Park, North Carolina 27709
September 1986
-------
September 29, 1986
NOTICE
The information described in this document has been funded as a part of
the National Acid Precipitation Assessment Program by the U. S. Environmental
Protection Agency.
This document has been subjected to the Agency's administrative review
and approved for distribution. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.
ii
-------
September 29, 1986
PREFACE
This document is a part of a larger overview document describing the
U. S. Environmental Protection Agency's Aquatic Effects Research Program.
The Eastern Lake Survey-Phase I (Synoptic Chemistry), described here, is the
first phase of one of several regionalized integrative studies being
conducted in areas of the United States in which surface waters are
potentially sensitive to change as a result of acidic deposition. The
objectives of these surveys are to quantify and/or define within these
potentially sensitive areas: (1) the present chemical status of surface
waters in the United States, (2) the temporal and spatial variability of
surface waters, (3) the biological resources associated with these surface
waters, (4) terrestrial factors controlling surface water response, and
(5) future trends in surface water chemistry and biology.
For technical information about the Eastern Lake Survey contact:
Dr. D. H. Landers, Technical Director
Eastern Lake Survey
U. S. Environmental Protection Agency
Environmental Research Laboratory
200 S.W. 35th Street
Con/all is, Oregon 97333
For information about other programs within the Aquatic Effects
Research Program, contact:
Dr. R. A. Linthurst, Director
Aquatic Effects Research Program
U. S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
Mail Drop - 39
Research Triangle Park, North Carolina 27711
iii
-------
September 29, 1986
CONTENTS
Page
Notice i i
Preface i i i
Figures v
Tab!es vi
Introduction 1
Approach 5
Sampling Design 5
Vari ables Selected for Analysi s 14
Analytical Methods 16
Quality Assurance Program 16
Implementation 24
Base Site Operations 24
Analytical Laboratory Operations 26
Resul ts 27
Number of Lakes Sampled 27
Population Estimates 27
Use of Eastern Lake Survey - Phase I Data by Other Projects 38
References : 39
iv
-------
September 29, 1986
FIGURES
Number Page
1 Geographic regions in which the Eastern Lake Survey -
Phase I was conducted 6
2 Northeastern subregions and alkalinity map classes,
Eastern Lake Survey - Phase I 8
3 Upper midwestern subregions and alkalinity map classes,
Eastern Lake Survey - Phase I 9
4 Southeastern subregions and alkalinity map classes,
Eastern Lake Survey - Phase I 10
5 Percentage of total samples analyzed for quality assurance
in the Eastern Lake Survey - Phase I 18
6 Data management approach for the Eastern Lake Survey -
Phase I 22
7 Cumulative frequency distributions for acid neutralizing
capacity for the population of lakes >4 ha and <2000 ha in
two regions and two subregions sampled in fall, 1984 during
Phase I of the Eastern Lake Survey 30
8 Cumulative frequency distributions for pH for the
population of lakes >4 ha and <2000 ha in two regions
and two subregions sampled in Tall, 1984 during Phase I of
the Eastern Lake Survey 32
-------
September 29, 1986
TABLES
Number Page
1 Eastern Lake Survey - Phase I Regions and Subregions 11
2 Summary of Variables Measured in the Eastern Lake Survey -
Phase I 15
3 Number of Lakes Sampled within Each Subregion during
the Eastern Lake Survey - Phase I 28
4 Number of Lakes Sampled within Each State during
the Eastern Lake Survey - Phase I 29
5 Estimated Total Number of Lakes (>4 ha and <2000 ha), and
Number and Percentage of Lakes with Selected Values of
pH and ANC from Phase I of the Eastern Lake Survey 33
6 Estimated Total Number of Lakes (>4 ha and <2000 ha), and
Number and Percentage of Lakes with Selected Values of Four
Key Variables from Phase I of the Eastern Lake Survey 34
vi
-------
September 29, 1986
INTRODUCTION
The first phase of the National Surface Water Survey (NSWS) was
implemented in fall 1984 with Phase I of the Eastern Lake Survey (ELS-I),
a synoptic chemical survey of 1,798 lakes in the eastern United States. The
NSWS, which includes both lakes and streams, is part of the National Acid
Precipitation Assessment Program (NAPAP). One of the responsibilities of
NAPAP is to assess the extent and severity of the risk that acidic
deposition poses to aquatic resources within the United States. The NSWS
was initiated by the U. S. Environmental Protection Agency (EPA) when it
became apparent that existing data could not be used quantitatively to
assess the present chemical and biological status of surface waters in the
United States.
The ELS-I was designed to reduce the uncertainty in making regional
scale assessments based on existing data by:
(1) providing data from a subset of lakes statistically selected
from the population of lakes within a region as a whole;
(2) using standardized methods to collect chemical data;
(3) measuring a complete set of variables thought to influence or
be influenced by surface water acidification;
(4) providing data that can be used statistically to investigate
relationships among chemical variables on a regional basis; and
(5) providing reliable estimates of the chemical status of lakes
within a region of Interest.
Environmental data collection activities, conducted or sponsored by
EPA, must be based on a program which ensures that the resulting data are of
known quality and are suitable for the purpose for which they are intended.
These requirements were incorporated into the ELS-I by: clearly defining
-------
September 29, 1986
ELS-I objectives; identifying intended uses and users of the data;
developing an overall conceptual and practical approach to meeting the
objectives; developing an appropriate survey design; identifying the
quality of data needed; developing analytical protocols and quality
assurance/quality control (QA/QC) procedures; testing the approach
through a "pilot" or feasibility study; and revising the approach or
methods as needed.
The ELS-I was designed to provide statistically comparable data that
could be extrapolated with a known degree of confidence to regional scales.
The primary objectives of the ELS-I were to determine:
(1) the percentage (by number and area) and location of lakes that are
acidic in potentially sensitive regions of the eastern United
States;
(2) the percentage (by number and area) and location of lakes that
have low acid neutralizing capacity in potentially sensitive
regions of the eastern United States; and
(3) the chemical characteristics of lake populations in potentially
sensitive regions of the eastern United States and to provide a
data base for selecting lakes for future studies.
The data generated by the Survey were intended to be used to assess the
present chemical status and extent of lakes potentially sensitive to acidic
deposition. The conceptual approach to the Survey emphasized that the data
would not be used to attribute observed conditions to acidic deposition or
to any other cause. Rather, the Survey's intent was to provide information
for development of correlative, not cause-and-effect, relationships.
Presumably, not all lakes have been affected by acidic deposition; this
suggests that the best approach to reducing the uncertainty relative to
regional scale effects is to identify and characterize the subset of lakes
that is sensitive to acidic deposition. Thus, the ELS-I was designed to
provide a geographically extensive data base of adequate quality to
estimate with known confidence the number of acidic and potentially
sensitive lakes, identify where they are located and describe their present
chemical status.
-------
September 29, 1986
The regional characterization approach developed for the ELS-I allows
many lakes in a large number of geographic regions to be characterized on
the basis of a subset of lakes within each region. In this synoptic
("snap-shot") approach, the minimum number of samples from each lake is
collected in order to maximize the number of lakes that can be sampled.
Because the chemistry of lakes varies seasonally, consideration of the
optimal period for sampling was a major part of the Survey design.
To extrapolate the data collected from a subset of systems to the total
lake population of interest requires that a statistically unbiased method of
lake selection be used. The lake selection procedure developed for the
ELS-I ensured that each lake to be sampled in a particular geographic region
had an equal probability of being included in the Survey.
The quality of the data needed to make regional scale assessments was
defined before the Survey was implemented. Because of the scope of the
project, eight base sites and four analytical laboratories were employed to
collect and analyze data. To ensure that data from each base site and
laboratory were comparable, standardized methods for sampling and analysis
1 2
were developed. The detailed quality assurance plan developed for the
Survey guided the collection of data and provided the basis for assessing
its quality in terms of accuracy, precision, bias, completeness and
comparability.
The complexity and scope of the ELS-I required that the project design
be tested before initiating the full-scale sampling effort. To identify
unforeseen problems in the conceptual approach, two feasibility ("pilot")
studies were conducted in the northeastern United States during January and
May, 1984. Sixty lakes were sampled during the winter pilot study and 113
during the spring pilot study. The objectives of both studies were to test
and evaluate sample site selection, sampling and analytical methods, field
base operations, sample shipping and tracking procedures, quality assurance
and quality control, and data management. Deficiencies identified in the
pilot studies were corrected for the full Survey.
Planning for the ELS-I began in October 1983. The research plan for
the NSWS was initially reviewed late in 1983 by 100 scientists with
A
expertise in various areas of study. Fifty scientists discussed the
-------
September 29, 1986
plan during a workshop held in December; and suggested modifications were
incorporated by March 1984. The research plan was submitted to members of
the American Statistical Association (ASA) for review in June 1984; a final
ASA review was conducted in October. The comments and responses from this
latter review appeared in the American Statistician.
The geographically extensive data base provided by the ELS-I serves as
the basis for future studies within the NSWS. Phase II is a study of the
temporal and spatial variability in lake water chemistry and will provide
the means to evaluate the degree to which the samples taken in Phase I
represent the chemical status of lakes during other seasons of the year.
The ELS-I data will also be used to develop the future Long-Term Monitoring
Project (LTM), designed to determine long-term trends in the chemical status
of surface waters; currently monitored sites as well as those sampled during
the NSWS will be evaluated for inclusion in this project. The approach used
in the ELS-I is summarized below.
-------
September 29, 1986
APPROACH
SAMPLING DESIGN
Identification of Study Area
Although the NSWS is a nationwide effort, many areas in the United
States are not likely to contain lakes that are presently acidic or at risk
of becoming so (primarily because of geologic factors that buffer acidic
inputs). Regions containing lakes that might be susceptible to change,
i.e., have minimal or low capacity to neutralize acidic inputs, were
delineated by Omernik and Powers using historical alkalinity data. Three
geographic regions (Figure 1) in the eastern United States were chosen for
study during the ELS-I: the Northeast (Region 1), the Upper Midwest (Region
2) and the Southeast (Region 3). These regions are estimated to contain
about 95 percent of the lakes in the eastern United States which have
alkalinity concentrations <400 ueq L" .
Lake Selection
Selection of the population of lakes for inclusion in the ELS-I was a
stepwise process that included:
(1) identifying homogeneous geographic areas (subregions) within the
three regions;
(2) differentiating, on the basis of alkalinity data, areas within
subregions expected to contain lakes within one of three
alkalinity classes (alkalinity map classes); and
(3) selecting lakes from each of the alkalinity map classes for
sampling, using a systematic procedure beginning with a random
starting point.
The terms alkalinity and acid neutralizing capacity (ANC) have been used
interchangeably to mean the amount of acid a surface water can assimilate
(or neutralize) before it becomes acidic. Alkalinity is a component of ANC
and, therefore, ANC is the more correctly used term to describe total
neutralizing capacity.
-------
September 29, 1986
Upper Midwest
Subregion Boundary
Figure 1. Geographic regions in which the Eastern Lake
Survey - Phase I was conducted.
-------
September 29, 1986
Because the statistical design used in lake selection has different levels
(or strata; steps 1 and 2) and because it employs a systematic selection
procedure (step 3), it is referred to as a "systematic, stratified design
with a random start." Each level of the design -- region, subregion, and
alkalinity map class -- is termed a "stratum."
Regions, subregions and alkalinity map classes (Figures 2-4) were
delineated on 1:250,000-scale U. S. Geological Survey (USGS) maps.
Boundaries of subregions were drawn on the basis of the similarities of
areas within regions with respect to surrounding physical geography,
land-use characteristics and available water quality information. The names
of the delineated subregions are listed in Table 1. Alkalinity map classes
789
were developed from existing alkalinity data. ' Three classes of
alkalinity were chosen: <100 ueq L , 100-200 ueq L and >200 ueq L"1.
These concentrations were chosen because there is evidence to suggest that
biological effects of acidification become apparent in the range of
10-90 ueq L" , and also because 200 ueq L has been used as the boundary
distinguishing lakes that are potentially sensitive to acidic deposition
from those that might be insensitive.
Each lake within the alkalinity map classes was systematically assigned
a unique identification number. The total number of lakes so designated
represented the "map population" of lakes. Because 1:250,000-scale USGS
maps inconsistently show lakes smaller than 4 hectares, these lakes
generally were not included in the map population or, consequently, in the
Survey. Lake identification numbers were entered into a computer file in
numerical order. The number of lakes within each alkalinity map class was
then divided by the number of lakes to be sampled to obtain a number (k). A
sample size of 50 was considered sufficient to characterize the population
of lakes within each alkalinity map class. The first sample lake was then
selected at random between 1 and k; thereafter, every kth lake was selected.
This procedure ensured that each lake within an alkalinity map class had an
equal probability of being selected for sampling; thus, this subset of lakes
is referred to as the "probability sample."
The selected lakes were examined on larger scale USGS maps to exclude
those that were of "non-interest" to the Survey or that proved to be
"non-lakes." Examples of these exclusion categories include flowing water,
-------
September 29, 1986
Alkalinity Map Classes
(^eq L")
T] < 100
T] 100-200
T] >200
— Subregion Boundary
Southern New England (1D)
Figure 2. Northeastern subregions and alkalinity map classes,
Eastern Lake Survey - Phase I.
8
-------
September 29, 1986
^>>-is=»>
Northeastern Minnesota (2A) 1
\ 3
^'2
Alkalinity Map Classes
U/eq L")
CD < 100
[Til 00-200
[T)>200
•— Subregion Boundary
Upper Great Lakes Area (2D)
Upper Peninsula of Michigan (2B)
Northcentral Wisconsin (2C)
Figure 3. Upper midwestern subregions and alkalinity map classes,
Eastern Lake Survey - Phase I.
-------
September 29, 1986
r
T
Southern Blue Ridge (3A)
Figure 4. Southeastern subregions and alkalinity map classes,
Eastern Lake Survey - Phase I.
10
-------
September 29, 1986
TABLE 1. EASTERN LAKE SURVEY - PHASE I REGIONS AND SUBREGIONS
Region
Subreglon
Code
Name
Code
Name
Northeast
Upper Midwest
Southeast
1A Adirondacks
IB Poconos/Catskills
1C Central New England
ID Southern New England
IE Maine
2A Northeastern Minnesota
2B Upper Peninsula of Michigan
2C Northcentral Wisconsin
2D Upper Great Lakes Area
3A Southern Blue Ridge
3B Florida
11
-------
September 29, 1986
marshes, and lakes surrounded by urban, industrial, or agricultural
development. To maintain the desired sample size of at least 50 lakes in
each alkalinity map class, the above process was repeated after eliminating
lakes in these exclusion categories. Additional exclusion categories were
identified during field sampling activities; examples of these included
sites that were visited but were found to be dry, or to be streams, and
lakes that were too shallow to permit collection of a sample free of debris
or sediment.
The lakes remaining after elimination by map examination and field
sampling activities comprised the "target population" of lakes. Some lakes
could not be visited due to weather conditions, denial of access permission,
or because they were frozen. Because these lakes could not be evaluated to
determine if they were in one of the exclusion categories, their target
status could not be determined and thus, they represent incompleteness in
the sample and cannot be included or excluded from the target population.
The equations developed for extrapolating the sample data to the target
population account for lakes of indeterminant target status.
The criteria developed during map examination and field sampling
activities, and other criteria developed during data analysis, collectively
identify the target population of lakes, which is smaller than the map
population. Conclusions based on the Survey data can be made only for those
lakes included in the target population. Lakes smaller than 4 hectares or
located in heavily industrialized areas, for example, cannot be
characterized with the ELS-I data. The flexibility of the sampling design
allows estimates to be made for any subset of lakes within the target
population (e.g., the number of shallow lakes or those with a particular pH
or sulfate concentration), provided the characteristics of that subset are
explicitly defined.
Because not all lakes within each stratum were sampled, equations were
developed to extrapolate the values measured in the sample lakes to obtain
estimated characteristics of the entire target population. The number of
lakes in the target population within each stratum differed; thus each lake
sampled represents a different number of lakes 1n the target population.
For example, 1f the target population size was estimated to be 500 and
50 lakes were sampled, each sampled lake represented 10 target lakes. For
12
-------
September 29, 1986
an estimated target population of 1500 lakes and a sample size of 50, each
sampled lake represented 30 lakes. Thus, obtaining estimated
characteristics for the entire target population in a region or subregion
requires that the sample data be adjusted or "weighted" to account for
different estimated target population sizes. These weights, which are
specific to each stratum, are applied to the sample values to obtain
estimated characteristics of all lakes in the target population.
Special Interest Lakes
Other lakes, in addition to those chosen with the statistical design,
were sampled during the ELS-I primarily because of their importance to
programs initiated before the NSWS. These "special interest" lakes included
those currently in the EPA Long-Term Monitoring Program, which also is a
part of NAPAP. Others were included on the basis of recommendations by the
National Research Council's Acid Deposition Trends Committee and by various
state and federal agencies. Because these lakes were not chosen by the
selection process discussed above, data collected from them were not used in
characterizing the target population of lakes. Special interest lakes are
thus differentiated from the probability sample lakes selected by the
statistical design.
Sampling Period
One of the primary goals of the ELS-I was to identify those lakes with
chemical characteristics that are "typical" of the total population of lakes
within a region. These lakes could then be considered for more detailed,
long-term study. To cover as much area as possible within the regions of
interest (to maximize spatial coverage) within one season (to minimize
seasonal variability) each lake was sampled in the fall. Fall was chosen as
the best season in which to sample because, as the upper water layers of
lakes begin to cool, temperature differences between the top and bottom are
minimized. As the water temperature becomes homogeneous throughout the
lake, the water layers begin to circulate and mix, eliminating most chemical
differences which may have previously existed. When this process, commonly
called "fall turnover," is completed, the variability of chemistry within
13
-------
September 29, 1986
any one lake is expected to be less than the variability between any two
lakes at that time. The fall sample thus serves as an "index" sample which
can be related to the chemical status during the winter, spring, and summer.
The relationship of fall chemistry to that of other seasons is being
evaluated in Phase II of the National Lake Survey.
VARIABLES SELECTED FOR ANALYSIS
Twenty-five physical and chemical variables (Table 2) were chosen for
analysis in the ELS-I. The variables were selected on the basis of their
importance in developing chemical characterization of lakes from a regional
perspective. The variables included those relating to the acid/base status
of lakes: pH, acid neutralizing capacity (ANC), and dissolved inorganic
carbon, which can be used to determine if the lake water is saturated with
respect to atmospheric carbon dioxide. Sulfate and nitrate were measured
because they are often the dominant acidic anions in acidic deposition.
Additional anions and such cations as calcium and magnesium were measured
not only to aid in chemically describing lakes but also to serve as a
quality assurance check on the methods used in their analyses (the sum of
cations theoretically equals that of anions).
A number of chemical variables important to biota were also included in
the analyses. These "biologically-relevant11 chemical variables include:
total phosphorus, dissolved silica, and ammonium, all of which are nutrient
sources. Additionally, total dissolved fluoride and dissolved organic
carbon can reduce the concentrations of potentially toxic forms of aluminum
and other metals that can be detrimental to organisms.
Some dissolved organic carbon compounds impart "true color" to lake
water; these organic acids can dissociate into organic ions and hydrogen ion
and thus contribute to acidity. Dissolved silica and total aluminum
concentrations also can be used as indicators of the amount of "mineral
weathering" in the watershed, i.e., dissolution of chemical compounds in the
surrounding soils. Weathering is one of several processes that can
contribute to the neutralization of acidic inputs.
14
-------
September 29, 1986
TABLE 2. SUMMARY OF VARIABLES MEASURED IN THE EASTERN LAKE SURVEY - PHASE Ia
Acid neutralizing capacity Magnesium, dissolved
Aluminum, extractable Manganese, dissolved
Aluminum, total Nitrate, dissolved
Ammonium, dissolved pH
Carbon, dissolved inorganic Phosphorus, total
Carbon, dissolved organic Potassium, dissolved
Calcium, dissolved Secchi disk transparency
Chloride, dissolved Silica, dissolved
Color, true Sodium, dissolved
Conductance Sulfate, dissolved
Fluoride, total dissolved Temperature
Iron, dissolved Turbidity
aThe complete list of variables for the Survey is given in
Linthurst et al., Overton et al., and Kanciruk et al.11>12<13
15
-------
September 29, 1986
ANALYTICAL METHODS
The analytical methods chosen for the NSWS were suitable for analyzing
samples from lakes expected to have low chemical concentrations.
Recommended methods could be used to detect concentrations at one-fifth to
one-tenth the values expected to occur in the sampled lakes. The methods
selected were state-of-the-art techniques and were extensively peer reviewed
before application in the Survey. A detailed description of analytical
o
techniques and procedures is provided by Hillman et al.
QUALITY ASSURANCE PROGRAM
An extensive quality assurance program designed to standardize all
sampling and analytical protocols and to ensure that the quality of the data
3 Id
could be determined was Implemented in the ELS-I.
Standardized Protocols
Field sampling methods used at the lake site, and field laboratory
activities are detailed in Morris et al. All personnel participating in
field activities were trained simultaneously. Site visits to each field
operations base were made throughout the sampling period to ensure that all
methods were being properly conducted. An extensive evaluation procedure
was used to select qualified analytical laboratories and to ensure their
continued acceptable performance throughout their participation in the
project.
Quality Assurance/Quality Control Samples
Quality control samples were analyzed during the Survey to ensure that
instruments and data collection activities were operating within the limits
of the QA plan. Examples of quality control samples include: duplicate
lake samples and deionized water blanks, and matrix spikes, which are
samples to which a known amount of the analyte was added, to determine
potential interferences in the analytical method. Quality assurance samples
included audit samples with known concentrations that were synthetically
prepared or collected in bulk from two lakes before the Survey was
implemented.
16
-------
September 29, 1986
The analytical laboratories were required to analyze samples within the
prescribed "holding times." Holding times are specific to each chemical
variable and are the maximum defined lengths of time that a sample can be
held after collection from a lake before analysis. Exceeding holding times
increases the risk of sample degradation. Results of analyses of quality
assurance samples were used to assess the amount of variability at each
level of activity among the field and analytical laboratories. More than
30 percent of the samples analyzed during the ELS-I were for QA/QC purposes
(Figure 5).
Data Base Quality Assurance
Three procedures to ensure the quality of the data base were
implemented during the ELS-I. Oak Ridge National Laboratory (ORNL),
responsible for data management, implemented quality assurance measures to
ensure that data entry was correct and that all subsequent modifications to
the data base were recorded and explicity documented. Data verification was
conducted by the EPA-Environmental Monitoring Systems Laboratory (EMSL) in
Las Vegas, Nevada. Data validation was the primary responsibility of the
EPA-Environmental Research Laboratory (ERL) in Corvallis, Oregon. Although
primary responsibility of each task was identified with a specific agency,
development of the final data base used in data analyses was an interactive,
collective effort. As a result of the quality assurance procedures, the
ELS-I data base contains four distinct data sets: raw, verified, validated,
and final. The data bases and the complete records of their development are
fully documented.
Data Verification--
Data verification was a systematic process established to review values
in the raw data set. To accomplish this task, the EMSL-Las Vegas Quality
Assurance Support Group worked closely with ORNL.
The initial step involved a review of the field data forms to ensure
that calibration and field quality control sample data were within
previously established acceptance criteria. Errors in transcription were
corrected. Redundant measures of pH and DIG in the field and analytical
laboratories, and the in situ lake pH determination, were also compared.
17
-------
CO
Duplicates
(9.7%)
Duplicates
(10.5%)
Blanks
(9.6%)
Northeast
Lab Audits
(2.7%)
m Field Audits
(5.4%)
Lake Samples
(72.6%)
Duplicates
(8.8%)
Blanks
(10.4%)
Lab Audits
(2.8%)
Field Audits
(5.5%)
Lake Samples
(70.8%)
Blanks
(8.3%)
Upper Midwest
Figure 5.
Lab Audits
(6.7%)
Field Audits
*" (5.9%)
Lake Samples
(70.3%)
Southeast
Percentage of total samples analyzed for quality assurance
in the Eastern Lake Survey - Phase I.
CD
3
CT
-------
September 29, 1986
The second step was to evaluate the results reported by the analytical
laboratories to ensure that the reports were complete, that internal quality
control criteria were met and that, if necessary, data were appropriately
qualified. Interaction with the analytical laboratories by the QA Support
Group involved obtaining confirmation or correction of reported data and,
when required, sample reanalysis.
After entry into the data base at ORNL, the data were then transferred
by tape to the National Computer Center in Research Triangle Park, North
Carolina from where they were accessed by the QA Support Group to implement
the third step in the verification procedure. The output from a series of
computer programs, the original data, and information in field notebooks
were then used to produce a standardized verification report. This report
included definitions of data qualifiers, documentation of data
resubmissions, and requests for reanalysis and confirmation, summaries of
modifications to the raw data base and an inventory of all verified samples.
Data were verified on a sample by sample basis using three types of
acceptance criteria: 1) individual samples were evaluated for internal
consistency such as anion-cation balance or conductance comparisons; 2) all
variables within a group of samples analyzed as a whole (batch) were
qualified if the results of the analysis of external QA samples such as
field blanks, duplicates or audits did not meet previously established
acceptance criteria; and 3) data were flagged if internal quality assurance
criteria such as matrix spike recovery, calibration and reagent blank
analysis, internal duplicate precision, required instrument detection limit,
percent recovery of quality control check samples, and holding time
requirements were not satisfied. The final verification step was the
transfer of information to ORNL so that the raw data base could be converted
into the verified data base.
Data Validation--
The validation process for the ELS-I data base was designed to
investigate potential errors in the chemical analyses that were not detected
in the verification step, and to provide a review of the quality of
non-chemical variables such as SeccM disk transparency, color, turbidity,
and physical measurements of the lake and watershed.15
19
-------
September 29, 1986
The data validation procedures provided the means to: 1) identify
questionable data based on empirical evidence or statistical analyses; and
2) determine the most appropriate value for a water quality variable when it
had been measured in more than one way or when data substitution was
necessary.
Two aspects of the validation approach were identification of possible
"outliers" and evaluation of systematic error in the measurements. Outliers
are values for variables that are not typical of other sample values
observed for the group of lakes from which samples are drawn.
Identification of outliers that may indicate random error in the data was
accomplished using various statistical procedures including univariate,
bivariate and multivariate analyses. An example of systematic error that
might result from an analytical methods problem would be that all values for
a variable, such as sulfate, were very low, indicating that the method for
measuring sulfate may have consistently underestimated the true value.
Methods for evaluating systematic error included comparing the ELS-I data to
data available from other investigations. These comparisons served to
identify data from the ELS-I which might require additional evaluation to
ensure their quality. The external data sets for comparison were selected
on the basis of their availability from the regions 1n which the ELS-I was
conducted, the accessibility to these data, and the availability of
accompanying documentation of quality assurance data.
Development of Final Data Set--
The final data set (Data Set 4) was prepared to resolve problems
relative to data interpretation because of missing values in the validated
data set. Data Set 4 also was modified by averaging field duplicate values
and substituting for analytical values determined to be in error during
validation.
Several substitution methods were used. When available, values from
duplicate samples were used. Dissolved organic carbon, pH and conductance
were determined using more than one analytical method; other variables were
measured on lake water samples that had been "split" or divided and analyzed
at separate laboratories. These "redundant" measurements were substituted
20
-------
September 29, 1986
for the missing value using a linear regression routine. If redundant
measurements were not available or acceptable, a substitution value was
calculated from the available data using observed relationships with other
variables (e.g., calcium and acid neutralizing capacity). The last option
used for identifying a substitution value was to use the stratum mean. All
substitution values were examined for acceptability before including them in
the final data set.
Two other changes were made in the final data set. If duplicate data
met QA precision criteria, the average value of the duplicates was used in
the final data set. Negative values for parameters other than acid
neutralizing capacity that resulted from analytical calibration bias were
set equal to zero. All values modified in the final data set were flagged
to indicate that they did not represent original measurement values.
Data Management--
The Survey has generated 1n excess of 1,000,000 data values. Because
of the size of the data base, stringent protocols to produce minimal errors
were Implemented. The Environmental Science Division of ORNL was
responsible for managing the ELS-I data base and for preliminary statistical
analyses. It served as the focal point for coordinating data flow
(Figure 6) from the QA Support Group at EMSL-Las Vegas to ERL-Corvallis.
The raw data set was produced by entering data directly from the field
forms and analytical laboratory forms. Data were entered independently by
two keyboard operators and automatically examined for data entry errors and
errors in range. A computer program was developed to compare the two sets
of entered data and to identify and correct inconsistencies.16
The data base was designed to permit the entry of data "tags",
characters which directly qualified the individual data value. Tags were
data qualifiers placed next to a variable on the data sheet by field
sampling crews or laboratory analysts. For example, if a pH meter was
unstable a V would be placed next to the recorded value by the technician.
Additionally, "flags" were used 1f the entered values exceeded
previously defined range limits or if pairs of variables (such as pH and
ANC values) within the same sample appeared Incompatible. Flags were
21
-------
September 29, 1986
SITE SELECTION ) ( FIELD SAMPLING ^ ( ANALYTICAL
V J \ ) \ LABORATORIES
VISUAL FORM CHECK
DATA ENTRY 1
DATA ENTRY 2
ERROR AND
RANGE CHECK
BATCH REPORTS
RAW DATA SET
(DATA SET 1)
VERIFIED DATA SET
(DATA SET 2)
DATA EDITING
AND FLAGGING
VALIDATED DATA SET
(DATA SET 3)
FINAL ELS-1 DATA SET
I DATA SET 4 )
Figure 6. Data management approach for the Eastern Lake Survey - Phase I.
22
-------
September 29, 1986
extensively used by EMSL-Las Vegas during the verification process. Thus,
the data base was structured so that each single variable value could have a
tag and/or flag if appropriate.
Data were transferred from the raw to the verified data set, and
subsequently, to the validated data set as the above procedures progressed.
As transfers were made, the original entries with all qualifiers were
maintained with backup files as a means of documenting all changes to the
data base.
23
-------
September 29, 1986
IMPLEMENTATION
Satisfying the data quality objectives that were established for the
ELS-I required that sample collection be conducted in a way that minimized
the risk of introducing variability. The large number of lakes and their
extensive geographic distribution, the relatively short optimal period for
sampling (fall turnover) and the previously established holding times for
critical chemical variables necessitated the use of a sampling procedure
that would allow samples from a large number of sites to be delivered as
quickly as possible to analytical laboratories for processing. For these
reasons, helicopters were chosen as the most effective means to access sites
and to deliver samples for processing. Because some of the critical
variables can change very quickly (within hours) after water samples are
removed from lakes, mobile field laboratories were used to measure some key
chemical constituents especially prone to change and to conduct preliminary
processing of the samples before delivery to analytical laboratories.
BASE SITE OPERATIONS
The sampling effort for the ELS-I began in October 1984 and ended in
December 1984. Eight bases of operations were established: Bangor, Maine;
Lake Placid/Saranac Lake, New York; Lexington, Massachusetts; Mt. Poconos,
Pennsylvania; Rhinelander, Wisconsin; Duluth, Minnesota; Asheville, North
Carolina; and Lakeland, Florida. Each base site consisted of a mobile field
laboratory, an area for storage and calibration of field equipment, two
helicopters, and a logistics coordination room. Ten personnel were
responsible for collection and delivery of samples to each field laboratory
which was staffed by a crew of five.
Sampling Activities
Approximately 12 lakes per day were sampled by two sampling crews whose
daily itineraries were Identified by logistics room staff. Watershed
descriptions for each lake visited were recorded upon approaching the lake.
24
-------
September 29, 1986
Temperature, pH, depth, conductance (a measure of dissolved chemicals) and
Secchi disk transparency (a measure of the depth of water clarity) were
determined on site. In all lakes greater than 1.5 meters in depth, samples
were taken from 1.5 meters below the surface. In shallow lakes, samples
were taken at depths less than 1.5 meters. Samples collected at each lake
included those to be processed through the field laboratory for delivery to
the analytical laboratories and two collected in sealed syringes carefully
protected from the atmosphere. Upon completion of each day's sampling
effort, lake and quality assurance samples (duplicates and deionized water
blanks) and data forms containing on-site records were delivered to the
field laboratory for processing and transmittal to the analytical
laboratories, EMSL-Las Vegas and ORNL, as appropriate.
Field sampling equipment was then subjected to various quality control
checks and prepared for the next day's sampling effort. Field crews joined
the logistics staff to report the day's activities and plan the next day's
itinerary.
Field Laboratory Activities
Upon delivery to the field laboratory, lake samples and blanks and
field laboratory QA samples were assigned identification numbers to assist
in sample tracking before field processing began. Special techniques were
used to analyze the duplicate syringe samples for pH and dissolved inorganic
carbon, which are especially prone to change as a result of exchange of
carbon dioxide with the atmosphere. Quick analysis and precautionary
measures taken with these unpreserved samples were necessary to prevent the
values from changing from those that actually occurred in the lake.
The remaining lake and QA samples were processed by preservation and/or
filtration; each was divided into seven portions (including one prepared for
analysis of extractable aluminum) and shipped by overnight courier in cooled
containers to one of four analytical laboratories. True color and turbidity
(a measure of particulate suspended particles) were also measured at the
field laboratory.
All field laboratory data were recorded on a standardized form and were
approved by the laboratory supervisor. This form was then transmitted with
the lake data form completed by the sampling crew, and the shipping form,
which was used to track all samples shipped from the laboratory.
25
-------
September 29, 1986
ANALYTICAL LABORATORY OPERATIONS
Four analytical laboratories processed approximately 2,650 samples
(both lake and QA) during the ELS-I: Global Geochemistry in Canoga
Park, California; Versar in Springfield, Virginia; EMSI (formerly Rockwell
International) in Newberry Park, California; and U.S. Geological Survey in
Denver, Colorado. All laboratories were contractually required to process
samples according to standardized protocols within pre-established holding
times. Data for each batch (an entire set of samples processed at one field
laboratory on one day) were required to be submitted within a period of 30
days to QA Managers at EMSL-Las Vegas.
26
-------
September 29, 1986
RESULTS
NUMBER OF LAKES SAMPLED
A total of 2,681 probability sample lakes were selected from the map
population for the ELS-I. Examination of 1:250,000-scale USGS maps resulted
in the elimination of 805 of these lakes. An additional 151 were classified
as non-target lakes by the field crews and 113 were not visited. Water
samples were collected from 1,612 lakes. Twenty of these lakes were not
included in the population estimates because they were larger than 2000 ha.
The final sample size, upon which population estimates were based, was
1,592.
A total of 763 probability sample lakes and 115 special interest lakes
were sampled in the Northeast (Table 3). Based on the sample size of
probability sample lakes, it is estimated that the number of lakes in the
Northeast characterized by the ELS-I is 7,096 with a standard error of
165.3. The estimated number of lakes characterized by the Survey in the
Upper Midwest was slightly larger (8,501) and in the Southern Blue Ridge
and Florida, smaller (258 and 2,098, respectively). The state in which the
largest number of probability sample lakes were sampled was Wisconsin
(253, Table 4), while the most special interest lakes sampled were in
New York (48).
POPULATION ESTIMATES
The ELS-I data were used to describe lakes in the eastern United States
by producing cumulative frequency distributions. Figure 7 is an example for
ANC. For any value of ANC shown on the x-axis, the corresponding percentage
of lakes estimated to have ANC equal to or less than that value is shown on
the y-axis. Because these distributions are estimated from sample data, the
95 percent confidence limit is also given (shown as a dashed line). For
example, if the percentage of lakes estimated to have ANC <0 ueq L"1 is 22.0
27
-------
September 29, 1986
TABLE 3. NUMBER OF LAKES SAMPLED WITHIN EACH SUBREGION
DURING THE EASTERN LAKE SURVEY - PHASE I
Estimated
Target Probability
Region
1: Northeast
2: Midwest
3: Southeast
Total
Subregion
1A: Adirondacks
IB: Poconos/Catskills
1C: Central New England
ID: Southern New England
IE: Maine
2A: Northeastern Minnesota
2B: Upper Peninsula of Michigan
2C: Northcentral Wisconsin
2D: Upper Great Lakes Area
3A: Southern Blue Ridge
3B: Florida
Population
Size0
1290 (47.6)
1479 (92.9)
1483 (57.5)
1318 (93.7)
1526 (66.0)
7096 (165.3)
1457 (74.3)
1050 (72.5)
1480 (48.8)
4515 (293.9)
8501 (315.5)
258 (20.5)
2098 (212.6)
Sample
Lakesb
155
143
163
126
176
763
147
146
153
141
587
94
148
1592
Special
Interest
Lakesc
48
12
49
0
6
115
9
10
32
1
52
10
9
186
Estimates are based on the number of probability sample lakes that were >4 ha and
<2000 ha. Standard errors on target population size estimates are shown in
parentheses.
The number of lakes sampled that were part of the probability sample.
cThe number of lakes sampled that were not part of the probability sample.
28
-------
September 29, 1986
TABLE 4. NUMBER OF LAKES SAMPLED WITHIN EACH STATE
DURING THE EASTERN LAKE SURVEY PHASE - I
Region State
1: Northeast Connecticut
Massachusetts
Maine
New Hampshire
New Jersey
New York
Pennsylvania
Rhode Island
Vermont
2: Midwest Michigan
Minnesota
Wisconsin
3: Southeast Florida
Georgia
North Carolina
South Carolina
Tennessee
Virginia
Probability
Sample
Lakes°
24
97
225
69
7
191
106
15
29
160
174
253
138
54
30
12
6
2
Special
Interest
Lakes b
0
0
6
17
12
48
0
0
32
11
10
31
9
1
7
0
2
0
aThe number of lakes (>4 ha and <2000 ha) actually sampled that were
part of the probability sample.
bThe number of lakes actually sampled that were not part of the
probability sample.
29
-------
September 29, 1986
80-
60-
40 -
20 -
80-
60
O
« 40
3
O.
O 20
° 80
S.
60
40
20
80
60
40
20
Northeast
Population Size = 7096(165.3)
Sample Size = 763
Upper Midwest
Population Size = 8501(315.5)
Sample Size = 587
Southern Blue Ridge .
Population Size = 258(20.5)
Sample Size s 94
Florida
Population Size * 2098(212.6)
Sample Size = 148
100 0 100 200300400500600700800900 1000
Acid Neutralizing Capacity (fieq t1)
Figure 7. Cumulative frequency distributions for acid neutralizing capacity for
the population of lakes >* ha and <2QOO ha in two regions and two
subregions sampled in fall, 1984 during Phase I of the Eastern Lake
Survey. The dashed line is the 95 percent upper confidence limit.
Population sise is estimated! standard errors of these estimates are
shown in parentheses. These plots can be used to make qualitative
comparisons among areas surveyed! e.g., the dots shown for the
Northeast indicate that approximately 30 percent of the lakes have
A1CC £100 JJeq L'l . A complete data report on the Survey results is
given in Linthurst et al., Overton et al., and Kanciruk et al. ' '
30
-------
September 29, 1986
and the 95 percent upper confidence limit on this estimate is 29.3, one
would be 95 percent certain that the true percentage of lakes with
ANC <0 ueq L is no greater than 29.3.
The cumulative frequency distributions for pH and ANC were distinctly
different among regions. Qualitative comparisons of the distributions of
ANC in the Northeast and Upper Midwest indicate that overall, the
northeastern lakes were characterized by lower ANC (Figure 7). Very few
lakes in the Southern Blue Ridge were characterized by low ANC although the
percentage of Florida (3B) lakes at the low end of the ANC range was
relatively high. Similar conclusions can be drawn by comparing the
distributions for pH (Figure 8).
Although these curves serve to highlight major differences among the
regional populations of lakes, they should not be used to provide
quantitative estimates. Instead, the equations developed in the design can
be used to generate quantitative estimates of characteristics of target
population lakes from sample data. For comparisons within and among
regions, any value of a chemical or physical variable of interest can be
selected. The values presented in Tables 5 and 6 were selected to quantify
the number of lakes (or the percentage of lakes) in the target population
that have a concentration equal to or less than, or equal to or greater
than, that listed.
Six of the chemical variables measured during the ELS-I (pH, ANC,
sulfate, extractable aluminum, dissolved organic carbon and calcium) were
selected for detailed analysis (Tables 5 and 6) because of their direct
relevance to the effects of acidic deposition on lake chemistry. In some
lakes, continuous inputs of acids can eventually result in decreases in pH
and acid neutralizing capacity. Lake waters can lose a high proportion of
their acid neutralizing capacity without experiencing substantial decreases
in pH; for this reason, losses in acid neutralizing capacity serve as a
better indicator of acidification than decreases in pH. Sulfate
concentrations in lake water can become elevated as a result of sulfate
deposition, one of the key components of acidic deposition. Some forms of
aluminum have been shown to increase with decreasing pH; thus, acidification
of lakes can be accompanied by elevated concentrations of aluminum that can
be toxic to aquatic organisms, particularly fish. Dissolved organic carbon
31
-------
September 29, 1986
o
fS
Q.
s.
* ha and <2000 ha In tvo regions and two subreglons sampled in fall,
198* during Phase I of the Eastern Lake Survey. Th« dashed lin* is the
95 percent upper confidence limit. Population sice is estljnatedi
standard errors of these estimates ara shovn in parentheses. These
plots can be used to make qualitative comparisons among araas surveyed.
Imprecise estimates can be obtained from the curvei e.g., the dots shovn
for tha Upper Midwest indicate that approximately 10 percent of the
lakes have pH <6.0. However, a complete description of the use of the
data base to make quantitative comparisons among areas is obtained in
the data report on the Survey results (Linthurst et al., Overton at al.,
11i12• 13
and Kanclruk «t aL.).
32
-------
TABLE 5. ESTIMATED TOTAL NUMBER OF LAKES (>4 HA AND <2000 HA), AND NUMBER AND PERCENTAGE OF LAKES WITH
SELECTED VALUES OF pH AND ANC FROM PHASE I OF THE EASTERN LAKE SURVEY. THE 95 PERCENT UPPER
CONFIDENCE LIMITS (UCL) FOR LAKE NUMBER ARE SHOWN IN PARENTHESES. THESE REFERENCE VALUES
PROVIDE ONE MECHANISM BY WHICH CHEMICAL CHARACTERIZATION OF LAKES WITHIN AND AMONG SUBREGIONS
CAN BE ACCOMPLISHED (Data are from Llnthurat et •!., Overt on et al., and Kanclruk et al.)
^
CJ
Reg ion/ Sub region
Northeast
Upper Mldveat
Southern Blue Ridge
Florida
oH ANC (ueq L~X)
Total <5.0 <6.0 <0 <200
Number
of Lakei Number (UCL) Percentage Number (UCL) Percentage Number (UCL) Percentage Number (UCL) Percentage
7096 240 (314) 3.4 916 (1056) 12.9 326 (422) 4.6 4258 (4513) 60.0
8502 130 (189) 1.5 818 (1036) 9.6 148 (209) 1.7 3518 (3982) 41.4
258 0 (-) 0 1 (2) 0.4 0 (-) 0 88 (108) 34.3
2098 259 (385) 12.4 687 (878) 32.7 463 (615) 22.0 1156 (1413) 55.1
Upper confidence limit* for value* of cero are undefined.
CO
rt>
TJ
CT
(D
-i
10
10
00
cr>
-------
TABLE 6. ESTIMATED TOTAL NUMBER OF LAKES (>4 HA AND <2000 HA), AND NUMBER AND PERCENTAGE OF LAKES WITH
SELECTED VALUES OF FOUR KEY VARIABLES FROM PHASE I OF THE EASTERN LAKE SURVEY. THE 95 PERCENT
UPPER CONFIDENCE LIMITS (UCL) FOR LAKE NUMBER ARE SHOWN IN PARENTHESES.* THESE REFERENCE VALUES
PROVIDE ONE MECHANISM BY WHICH CHEMICAL CHARACTERIZATION OF LAKES WITHIN AND AMONG SUBREGIONS
11 12,13
CAN BE ACCOMPLISHED (Data ace from Linthurst et al., Overton et al., and Kanciruk et al.)
Reglon/Subreglon
Northeast
Upper Midwest
Southern Blue Ridge
Florida
Sulfate Extractable Al Dissolved Organic Carbon Calcium
Total > 150 ueq L > 150 ug L"1^ > 6 ueq L"1 < 50 ueq L"1
of Lakes Number (UCL) Percentage Number (UCL) Percentage Number (UCL) Percentage Number (UCL) Percentage
7096 1846 (2082) 26.0 92 (135) 2.0 1873 (2090) 26.4 359 (442) 5.1
8502 608 (932) 7.1 2 (4) 0.0 5351 (5938) 62.9 776 (997) 9.1
258 22 (36) 8.5 0 (0) (-) 16 (28) 6.1 31 (42) 12.0
2098 846 (1088) 40.3 14 (35) 1.5 1445 (1766) 68.9 402 (551) 19.2
Upper confidence limits for values of zero are undefined.
Data are for "clearwater" lakes only. I.e., with true color values £30 platinum cobalt units.
ID
TJ
e+
(D
CT
fD
-J
ro
vo
CO
en
-------
September 29, 1986
(DOC) in "colored" lakes is largely composed of organic acids that are of
terrestrial origin; these compounds can complex, i.e., combine with, certain
types of ions, including toxic metals, and reduce their potential effects on
aquatic organisms. Additionally, DOC compounds can serve as sources of
hydrogen ion (i.e., acidity). Thus, some lakes that are acidic may be so
because of the presence of organic acids and not necessarily because of
acidic deposition. In many lakes, the dominant cation is calcium which Is
often equivalent in concentration to bicarbonate, a major component of ANC.
Thus, in lakes with low ANC, calcium may also be low, and may provide
evidence that the lake is potentially sensitive to acidic deposition.
In the Northeast, an estimated 240 lakes had pH <5.0 and 916 had
pH <6.0 (Table 5). In Florida a similar number of lakes (259) were
estimated to have pH <5.0 and substantially fewers (687) lakes had pH <6.0.
Acid neutralizing capacity values <0 ueq L"1 indicate that the lake
water has no capacity to neutralize acidic inputs. By definition, lakes
with ANC <0 ueq L"1 are acidic. The overall results for the Northeast and
Upper Midwest indicate that less than 5 percent and 2 percent, respectively,
of the lakes were acidic (Table 5). The majority of these acidic lakes were
located in a single subregion in each area; 10.7 percent of the lakes in the
Adirondacks (1A) were acidic, and 9.8 percent in the Upper Peninsula of
Michigan (2B).11 In Florida, 22 percent of the lakes were acidic, while in
the Southern Blue Ridge, no acidic lakes were sampled. The Northeast had
the highest number and percentage of lakes with ANC <200 ueq L"1 (4,258
and 60.0%, respectively).
An estimated 1,846 lakes of the total estimated target population
(7,096) in the Northeast had high sulfate concentrations (>150 ueq L"1;
Table 6). Of the five northeastern subregions surveyed, the highest numbers
of lakes with high sulfate concentrations occurred in the Poconos/
Catskills (IB) and Southern New England (ID).11 Fewer lakes with high
sulfate were found in the Upper Midwest and these were concentrated in the
Upper Great Lakes Area (2D). Only 22 lakes in the Southern Blue Ridge (3A)
were estimated to have sulfate concentrations >150 ueq L . In Florida
40.3 percent of the lakes in the target population were estimated to have
sulfate concentrations >150 ueq L .
35
-------
September 29, 1986
Extractable aluminum is an operationally defined term used to describe
forms of aluminum that can be toxic to fish and other aquatic organisms.
Concentrations of extractable aluminum were relatively low in all areas
surveyed. In the Northeast, 92 "clearwater" (true color <30 platinum cobalt
units) lakes had high (>150 ug L~ ) extractable aluminum, and of these, 82
were located in the Adirondacks (Table 6). The only other subregion where
more than one percent of the lakes were estimated to have high extractable
aluminum was Florida (3B) where 14 lakes representing 1.5 percent of the
target population had concentrations >150 ug L .
On a regional basis, a substantial percentage of lakes contained DOC
concentrations >6 mg L (Table 6). The Upper Midwest contained, by far,
the largest number of high DOC lakes while the number in the Southern Blue
Ridge was very small. Of the estimated 1,873 lakes in the Northeast with
high DOC, the highest number (643) was located in Maine (IE).11 In the
Upper Midwest, the highest percentages were observed in Northeastern
Minnesota (2A, the subregion that contains the Boundary Waters Canoe Area)
and the Upper Great Lakes Area (2D). In Florida (3B) 1,445 lakes,
representing an estimated 68.9 percent of the target population had high
DOC.
On a regional basis, Florida (3B) had the highest estimated
percentage (19.2%) of lakes with calcium concentrations <50 ueq L"1
(Table 6). The largest number of lakes with low calcium was located in the
Upper Midwest; of these, 324 were located in Northcentral Wisconsin (2C),
which comprised an estimated 21.9 percent of the target population. In the
Northeast, the highest percentage of lakes with low calcium was observed in
Southern New England (ID, 10.1%) followed by the Adirondacks (1A, 8.3%).
Because the ELS-I was not designed to evaluate cause-and-effect
relationships without additional supporting analyses, acidic deposition
cannot be inferred to be the cause of the above chemical observations.
However, the extensive data base generated by the ELS-I can be used in
conjunction with additional data to increase the confidence in the
hypothesized relationships and mechanisms believed to result from, or to be
affected by, acidic deposition. Examples of the questions generated from
correlative analysis of the data base that can be examined during future
studies include:
36
-------
September 29, 1986
• What Is the relationship between the observed sulfate
concentrations in lakes and the atmospheric deposition
of sulfate?
• How important are organic acids in explaining the
occurrence of acidic lakes?
• Can the acidic lakes in the Adirondacks, the Upper Peninsula of
Michigan, Northcentral Wisconsin and Florida be attributed to
acidic deposition?
• How important are acidic inputs derived from acidic
deposition in explaining the observations that lakes
with the highest extractable aluminum concentrations
occurred in the Adirondacks?
37
-------
September 29, 1986
USE OF EASTERN LAKE SURVEY - PHASE I DATA BY OTHER PROJECTS
Within EPA's Aquatic Effects Research Program, a number of projects use
the data base generated by the ELS-I. One hundred and fifty lakes were
selected from the lakes sampled during the ELS-I in the Northeast for
Phase II of the National Lake Survey. This component of the NSWS is
designed to quantify temporal variability in lake chemistry and to determine
present biological status in lakes.
Sites for the Direct/Delayed Response Project in the Northeast were
also selected from lakes sampled in the ELS-I. In this project, soils from
145 watersheds were sampled to obtain data that can be used in models
designed to predict the time required for surface waters to become acidic.
The ELS-I data are also being used to select lakes that can be
evaluated for inclusion in the future Long-Term Monitoring Project, designed
to examine trends in the response of surface waters to acidic deposition.
38
-------
September 29, 1986
REFERENCES
1. Morris, F. A., D. V. Peck, M. B. Bonoff and K. J. Cabbie. 1986.
National Surface Water Survey, Eastern Lake Survey - Phase I, Field
Operations Report. EPA/600/4-86/010, U. S. Environmental Protection
Agency, Las Vegas, Nevada.
2. Hillman, D. C. J., J. F. Potter and S. J. Simon. 1986. National
Surface Water Survey, Eastern Lake Survey - Phase I, Analytical Methods
Manual. EPA/600/4-86/009, U. S. Environmental Protection Agency,
Las Vegas, Nevada.
3. Drouse, S.K., D.C.J. Hillman, L.W. Creelman, J.F. Potter and
S.J. Simon. 1986. National Surface Water Survey - Phase I,
Eastern Lake Survey, Quality Assurance Plan. EPA/600/4-86/008,
U. S. Environmental Protection Agency, Las Vegas, Nevada.
4. Anonymous. 1984. National Surface Water Survey, National Lake
Survey - Phase I, Research Plan. U. S. Environmental Protection
Agency, Washington, D.C. (internal document).
5. Linthurst, R. A. and W. S. Overton. 1985. ASA review of EPA funded
acid precipitation research: Response to ASA Coordinating Committee's
comments on Aquatics Research Project 3B: National Surface Water
Survey: National Lake Survey - Phase I Research Plan. Amer. Stat.
39:266-271.
6. Omernik, J.M. and C.F. Powers. 1983. Total alkalinity of surface
waters—a national map. Ann. Assoc. Am. Geog. 73:133-136.
7. Omernik, J.M. and A.J. Kinney. 1986. Total Alkalinity of Surface
Waters: A Map of the New England and New York Region.
EPA-600/D-84-216, U. S. Environmental Protection Agency, Corvallis,
Oregon.
8. Omernik, J.M. and G.E. Griffith. 1985. Total Alkalinity of Surface
Waters: A Map of the Upper Midwest Region. EPA/600/D-85/043,
U. S. Environmental Protection Agency, Corvallis, Oregon.
9. Omernik, J.M. 1985. Total Alkalinity of Surface Waters: A Map of the
Appalachian Region. U. S. Environmental Protection Agency, Corvallis,
Oregon (draft).
10. Altshuller, A.P. and R.A. Linthurst (eds.) 1984. The Acidic Deposition
Phenomenon and Its Effects. Critical Assessment Review Papers,
Vol. II: Effects Sciences. EPA/600/8-83/016 BF, U. S. Environmental
Protection Agency, Washington, D.C.
39
-------
September 29, 1986
11. Linthurst, R.A., D.H. Landers, J.M. Eilers, D.F. Brakke, W.S. Overton
E.P. Meier, and R.E. Crowe. 1986. Characteristics of Lakes in the
Eastern United States: Volume I. Population Descriptions and
Physico-Chemical Relationships. EPA/600/4-86/007a, U. S. Environmental
Protection Agency, Washington, D.C. 139 pp.
12. Overton, W.S., P. Kanciruk, L.A. Hook, J.M. Eilers, D.H. Landers,
D.F. Brakke, D.J. Blick, Jr., R.A. Linthurst, M.D. DeHaan, and
J.M. Omernik. 1986. Characteristics of Lakes in the Eastern United
States: Volume II. Lakes Sampled and Descriptive Statistics for
Physical and Chemical Variables. EPA/600/4-86/007b,
U. S. Environmental Protection Agency, Washington, D.C. 374 pp.
13. Kanciruk, P., J.M. Eilers, R.A. McCord, D.H. Landers, D.F. Brakke, and
R.A. Linthurst. 1986. Characteristics of Lakes in the Eastern United
States: Volume III. Data Compendium of Site Characteristics and
Chemical Variables. EPA/600-4-86/007c, U. S. Environmental Protection
Agency, Washington, D.C. 439 pp.
14. Best, M.D., L.W. Creelman, S.K. Drouse, and D.J. Chaloud. 1986.
National Surface Water Survey, Eastern Lake Survey - Phase I, Quality
Assurance Report. EPA/600/4-86/011, U. S. Environmental Protection
Agency, Las Vegas, Nevada.
15. Eilers, J.M., D.J. Blick, Jr., and M.D. DeHaan. 1986. National
Surface Water Survey, Eastern Lake Survey - Phase I, Validation of the
Eastern Lake Survey - Phase I data base. U. S. Environmental
Protection Agency, Corvallis, Oregon.
16. Rosen, A.E. and P. Kanciruk. 1985. A generic data entry quality
assurance tool. Proceedings of the Tenth Annual SAS (Statistical
Analysis System, Inc.) Users Group International Conference.
March 10-13. Reno, Nevada.
40
------- |