United States
Environmental Protection
Agency
Office of Research
and Development
Washington DC 20460
EPA/600/4-86/007a
June 1986
vvEPA
Acid Deposition Aquatic Effects Research Program
Characteristics of
Lakes in the Eastern
United States
Volume I. Population
Descriptions and
Physico-Chemical
Relationships
-------
Upper Midwest
Southern New England (1D)
Northeast
Northeastern Minnesota (2A)
Upper Peninsula of Michigan (2B)
Adirondacks(IA)
Northcentral Wisconsin (2C)
Upper Great Lakes Area (2D)
Central New England (1C)
Poconos/Catskills(1B)
/
ern Blue Ridge (3A)
Southeast
Regions and Subregions, Eastern Lake Survey-Phase I
-------
EPA/600/4-86/007a
June 1986
Characteristics of Lakes in the
Eastern United States
Volume I. Population Descriptions and
Physico-Chemical Relationships
A Contribution to the
National Acid Precipitation Assessment Program
U.S. Environment?! Protection Agency
Region V, Library
230 South DsarLorn Street ^
Chicago, Illinois €0604
U.S. Environmental Protection Agency
Office of Research and Development, Washington, DC 20460
Environmental Research Laboratory, Corvallis, Oregon 97333
Environmental Monitoring Systems Laboratory, Las Vegas, Nevada 89109
-------
Notice
The information in this document has been funded wholly or in part by the U.S.
Environmental Protection Agency under Contract No. 68-03-3249 and 68-03-
3050 to Lockheed Engineering and Management Services Company, Inc., No.
68-02-3889 to Radian Corporation, No. 68-03-3246 to Northrop Services, Inc.,
and Interagency Agreement No. 40-1441-84 with the U.S. Department of
Energy. It has been subject to the Agency's peer and administrative review, and
it has been approved for publication as an EPA document.
Mention of corporation names, trade names or commercial products does not
constitute endorsement or recommendation for use.
Proper citation of this document is as follows:
Linthurst, R. A.1, D. H. Landers2, J. M. Eilers3, D. F. Brakke4, W. S. Overton8, E. P.
Meier6, and R. E. Crowe6. 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, DC,
1986,136pp.
Overton, W. S., P. Kanciruk7, L A. Hook8, J. M. Eilers, D. H. Landers, D. F. Brakke,
D. J\ Blick, Jr.3, R. A. Linthurst, M. D. DeHaan3, and J. M. Omernik9.
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, DC, 1986, 374
pp.
Kanciruk, P., J. M. Eilers, R. A. McCord8, D. H. Landers, D. F. Brakke, and R. A.
Linthurst. Characteristics of Lakes in the Eastern United States. Volume III. Data
Compendium of Site Characteristics and Chemical Variables. EPA/600/4-
86/007 c, U.S. Environmental Protection Agency, Washington, DC, 1986, 439
PP-
'USEPA Office of Research and Development, 401 M Street, S.W., Washington, DC 20460. Present address:
USEPA Environmental Monitoring Systems Laboratory, Mail-Drop 39, Research Triangle Park, North Carolina
27711.
'State University of New York, State University Research Center at Oswego, 300 Washington Blvd., Oswego, New
York 13126. Present address: USEPA Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis,
Oregon 97333.
3Northrop Services, Inc., USEPA Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis, Oregon
97333.
'Western Washington University, Institute for Watershed Studies, Environmental Sciences Bldg., Room 600,
Bellingham, Washington 98225.
'Oregon State University, Department of Statistics, Kidder No. 8, Corvallis, Oregon 97331.
6USEPA Environmental Monitoring Systems Laboratory, 944 E. Harmon Avenue, Las Vegas, Nevada 89114.
'Environmental Sciences Division, Oak Ridge National Laboratory, Post Office Box X, Oak Ridge, Tennessee
37831. Operated by Martin Marietta Energy Systems, Inc., under Contract No. DE-AC05-840R21400 for the U.S.
Department of Energy.
"Science Applications International Corporation, 800 Oak Ridge Turnpike, Oak Ridge, Tennessee 37831.
'USEPA Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis, Oregon 97333.
Inquiries regarding the availability of the Eastern Lake Survey-Phase I data base should be directed, in writing, to:
Chief, Air Branch, USEPA Environmental Research Laboratory, 200 S.W. 35th Street, Corvallis, Oregon 97333.
Environment-!
-------
Preface
As part of the National Acid Precipitation Assessment Program, a Federal
Interagency Task Force mandated by Congress in 1§S(L the United States
Environmental Protection Agency (EPA) initiated the National Surface Water
Survey (NSWS). The purposes of the NSWS are to assess the present chemistry
of surface waters, quantify the temporal variability and key biological resources
associated with these surface waters and initiate long-term monitoring in
characteristic systems. The NSWS is a three-phase study focusing on regions
of the U.S. that are potentially susceptible to change as a result of acidic deposition.
y^"1
The NSWS is one of the several major projects in the Acid Deposition Aquatic
Effects Research Program. This program, one of many research programs
addressing acidic deposition, is administered in the Acid Deposition and
Atmospheric Research Division: Office of Acid Deposition, Environmental
Monitoring and Quality Assurance in the EPA Office of Research and
Development.
The Aquatic Effects Research Program addresses four primary policy-related
questions:
1. How extensive is the damage to aquatic resources as a result of current
levels of acidic deposition?
2. What is the anticipated extent and rate of change to these resources in the
future?
3. What levels of damage to sensitive surface waters are associated with
various rates of acidic deposition?
4. What is the rate of change or recovery of affected systems, given decreases
in acidic deposition rates?
Fou/ major research projects within the Aquatic Effects Research Program
specifically address these policy questions within a regionalized framework.
These projects and their goals are:
(1) National Surface Water Survey (NSWS): to determine the present chemistry,
characterize the temporal variability in chemistry, and determine the key
biological resources of lakes and streams in potentially sensitive regions of
the U.S.;
(2) Direct/Delayed Response Project: to predict future changes in these
resources at present levels of acidic deposition, giving consideration to both
the terrestrial and aquatic variables that influence these changes;
(3) Watershed Manipulation Project: lo verify that predictions of future change
are reasonably sound by manipulating watershed catchments or system
components; and
(4) Long-Term Monitoring Project: to test the validity of predicted future
changes through long-term monitoring of regionally-characteristic lake
and stream systems.
The NSWS, including surveys of both lakes and streams, addresses the first goal
of the Aquatic Effects Research Program. The Eastern Lake Survey-Phase I
(ELS-I) was designed to statistically describe present surface water chemistry on
a regional scale.
Hi
-------
To further the current understanding of the effects of acidic deposition on
aquatic resources requires that the present chemical status of surf ace waters be
understood on large geographical scales. As individuals within the program and
others in the scientific community continue to analyze the ELS-I data, we
anticipate important scientific contributions to the understanding of regional
lake chemistry and its relationship to atmospheric deposition. Causality cannot
be determined from the results of ELS-I alone. Determining the relationships
between acidic deposition and lake chemistry are the goals of future projects
within the Aquatic Effects Research Program.
IV
-------
Volume I
Contents
Section Page
Notice ii
Preface iii
Volume II Contents ix
Volume III Contents x
Figures xi
Tables xv
Related Documents xix
Contributors xx
Acknowledgments xxii
Executive Summary xxv
1. Introduction 1
2. Methods 3
2.1 Design 3
2.1.1 Data Quality Objectives 3
2.1.2 Lake Representation 3
2.1.3 Statistical Design 3
2.1.4 Identification of Study Area 4
2.2 Lake Selection 4
2.2.1 Delineation of Strata 4
2.2.2 Development of the Map Population and Frame 4
2.2.3 Probability Sample 6
2.2.4 Identification and Refinement of Frame Population 6
2.2.5 Special Interest Lakes 9
2.2.6 Final Lake Lists and Maps 9
2.3 Applications of the Design 10
2.3.1 Defining the Target Population 10
2.3.2 Estimating the Target Population Size and
Associated Variance from Sample Data 11
2.3.3 Subpopulations 12
2.3.4 Restrictions 13
2.4 Lake Characterization 13
2.4.1 Lake Area and Elevation 13
2.4.2 Watershed Area 13
2.4.3 Lake Type 13
2.5 Water Sample Parameters and Analytical Methodology 14
2.6 Field Methods 14
2.6.1 Site Description 17
2.6.2 In Situ Measurements 17
2.6.3 Collection of Water Samples 17
2.6.4 Field Laboratory Activities 17
-------
Volume I
Contents (continued)
Section Page
2.7 Analytical Support 20
2.8 Data Base Management 20
2.8.1 Overview 20
2.8.2 Data Base Design and Data Flow 20
2.8.3 Data Base Structure 22
3. Quality Assurance 23
3.1 Preparation for the Survey 23
3.2 Implementation of Quality Assurance/Quality Control
Activities 23
3.3 Data Verification and Validation 23
3.3.1 Verification 24
3.3.2 Validation 26
3.4 Development of Final Data Set 27
3.5 Quality Assurance/Quality Control Results 28
3.5.1 Site Confirmation and Characterization 28
3.5.2 Evaluation of QA Sample Data 28
4. Results of Population Estimates 32
4.1 Data Presentation and Considerations 32
4.1.1 Presentation 32
4.1.2 Design Considerations 32
4.1.2.1 Design Constraints 32
4.1.2.2 Data Quality 33
4.2 Description of the Target Population 33
4.2.1 Number of Lakes Samples 33
4.2.2 Treatment of Large, Shallow and Thermally-
Stratified Lakes 35
4.2.3 Target Population and Population Estimates 35
4.3 Descriptive Statistics and Cumulative Distributions 36
4.3.1 Population Distributions 36
4.3.2 Definitions of Descriptive Statistics:
Interpretation of F(x) and G(x) Data Output 36
4.3.2.1 Distributions 45
4.3.2.2 Statistics 45
4.3.3 Comparisons of Distributions 46
4.3.4 Interpretation of Alkalinity Map Classes 46
4.4 Physical Characteristics of Regional Lake Populations 49
4.4.1 Northeast 49
4.4.2 Upper Midwest 49
4.4.3 Southeast 49
4.4.3.1 Southern Blue Ridge—Subregion 3A 49
4.4.3.2 Florida—Subregion 3B 50
4.5 Regional, Subregional and State Population Estimates:
ANC and pH 50
VI
-------
Volume I
Contents (continued)
Section Page
4.5.1 Acid Neutralizing Capacity 50
4.5.1.1 Reference Values 50
4.5.1.2 Northeast 50
4.5.1.3 Upper Midwest 51
4.5.1.4 Southeast 52
4.5.1.5 Estimates by State for ANC 52
4.5.1.6 Stratified and Shallow Lakes 53
4.5.2 pH I.. 55
4.5.2.1 Reference Values 55
4.5.2.2 Northeast 55
4.5.2.3 Upper Midwest 55
4.5.2.4 Southeast 55
4.5.2.5 Estimates by State for pH 55
4.6 Regional and Subregional Population Estimates for Other
Primary Variables 56
4.6.1 Sulfate 56
4.6.2 Calcium 58
4.6.3 Extractable Aluminum 60
4.6.4 Dissolved Organic Carbon 60
4.7 Statistics for Population Descriptions, Primary Variables 62
4.8 Statistics for Population Descriptions, Secondary
Variables 62
4.8.1 Nutrients 62
4.8.2 True Color, Turbidity and Secchi Disk Transparency 63
4.8.3 Sodium, Potassium and Magnesium 63
4.8.4 Iron, Manganese and Total Aluminum 64
4.8.5 Other Secondary Variables 65
4.9 Characteristics of Special Interest Lakes 65
5. Results and Discussion of Associations Among Variables 67
5.1 Relationships of pH and ANC — 67
5.1.1 Comparison of pH Measurements 67
5.1.2 Relationship between pH and ANC 68
5.1.3 Comparison of Lake Sensitivity Indices 72
5.2 Selected Associations Among Chemical Variables 74
5.2.1 Introduction 74
5.2.2 Sulfate 75
5.2.3 Extractable Aluminum 75
5.2.3.1 Background 75
5.2.3.2 Associations between Extractable
Aluminum and Other Variables 77
5.2.4 ANC versus Base Cations 79
5.2.5 Major Cations and Anions 84
5.2.5.1 Individual Example Lakes 84
5.2.5.2 Order of Major Ions 86
5.2.5.3 Relationships among Major Ions 91
vfi
-------
Volume I
Contents (continued)
Section Page
5.2.6 Dissolved Organic Carbon 94
5.2.6.1 Color and DOC 94
5.2.6.2 Anion Deficit 97
5.3 Hydrology 100
5.3.1 Hydrologic Lake Type 100
5.3.2 Hydraulic Residence Time 103
5.4 Characteristics of Acidic Lakes 108
6. Regional and Subregional Characteristics 110
6.1 Northeast 110
6.1.1 Adirondacks (1 A) 111
6.1.2 Poconos/Catskills (1B) 112
6.1.3 Central New England (1C) 112
6.1.4 Southern New England (1D) 113
6.1.5 Maine(1E) 113
6.2 Upper Midwest 114
6.2.1 Northeastern Minnesota (2A) 114
6.2.2 Upper Peninsula of Michigan (28) 115
6.2.3 Northcentral Wisconsin (2C) 115
6.2.4 Upper Great Lakes Area (2D) 115
6.3 Southeast 116
6.3.1 Southern Blue Ridge (3A) 116
6.3.2 Florida (3B) 116
7. Summary Observations 119
7.1 Objectives 119
7.2 Extent and Location of Acidic and Low pH Lakes 119
7.2.1 Acidic Lakes 119
7.2.2 Low pH Lakes 119
7.3 Extent and Location of Low ANC Lakes 120
7.4 Chemical Characterization 120
7.4.1 Sulfate 120
7.4.2 Calcium 120
7.4.3 Extractable Aluminum 120
7.4.4 Dissolved Organic Carbon 121
7.4.5 Major Cations and Anions 121
7.5 Future Studies 121
8. References 123
9. Glossary 130
VIII
-------
Volume II
Contents
Section Page
Notice ii
Related Documents iv
Volume I Contents v
Volume III Contents xii
Figures xiii
Tables xxiii
1. Introduction 1
2. Descriptions/Definitions of Parameters 4
3. Maps Showing Lake Locations 11
4. Regular Lakes Sorted by Lake ID 34
5. Regular Lakes Sorted by State and Lake Name 89
6. Special Interest Lakes Sorted by Lake ID 144
7. Special Interest Lakes Sorted by State and Lake Name 152
8. USGS Topographic Maps 160
8.1 Small-Scale Maps 160
8.2 Large-Scale Maps 160
9. Population Estimates for Selected Physical and Chemical Variables .. 187
9.1 Applications of the Design 187
9.1.1 Defining the Target Population 187
9.1.2 Estimating the Target Population Size and Associated
Variance from Sample Data 188
9.1.3 Subpopulations 190
9.1.4 Restrictions 192
9.2 Descriptive Statistics and Cumulative Distributions 192
9.2.1 Population Distributions 193
9.2.2 Definitions of the Descriptive Statistics:
Interpretation of F(x) and G(x) Data Output 193
9.2.2.1 Distributions 194
9.2.2.2 Statistics 195
ix
-------
Volume III
Contents
Section Page
Notice ii
Related Documents iv
Volume I Contents v
Volume II Contents xi
1. Introduction 1
2. Descriptions/Definitions of Parameters 3
3. Data for Individual Regular Lakes Sorted by Lake ID 13
4. Data for Individual Special Interest Lakes Sorted by Lake ID 337
Appendices
Appendix A 377
Appendix B 432
-------
Figures
Number Page
2-1 Regions and subregions surveyed during the Eastern Lake
Survey-Phase I 5
2-2 Northeastern subregions and alkalinity map classes, Eastern
Lake Survey-Phase I 6
2-3 Upper midwestern subregions and alkalinity map classes,
Eastern Lake Survey-Phase 1 7
2-4 Southeastern subregions and alkalinity map classes, Eastern
Lake Survey-Phase I 8
2-5 Estimating the target population size. Eastern Lake Survey-
Phase I 11
2-6 Field sampling activities. Eastern Lake Survey-Phase I 16
2-7 Field laboratory activities, Easten Lake Survey-Phase I 18
2-8 Data base development. Eastern Lake Survey-Phase I 21
3-1 Collection and processing of QA and QC samples, Eastern Lake
Survey-Phase I 24
3-2 Data verification procedures. Eastern Lake Survey-Phase I 25
3-3 Data validation procedures. Eastern Lake Survey-Phase I 26
3-4 Development of Data Set 4, Eastern Lake Survey-Phase 1 27
4-1 F(x) and G(x) distributions of ANC (/ueq L~1) for the target
population of lakes (<2000 ha) sampled in Region 1
(Northeast), Eastern Lake Survey-Phase I 37
4-2 F(x) and G(x) distributions of ANC {/ueq L"1) for the target
population of lakes (<2000 ha) sampled in Region 2 (Upper
Midwest), Eastern Lake Survey-Phase I 38
4-3 F(x) and G(x) distributions of ANC (/ueq L~1) for the target
population of lakes (<2000 ha) sampled in Subregion 3A
(Southern Blue Ridge), Eastern Lake Survey-Phase I 39
4-4 F(x) and G(x) distributions of ANC (/ueq L~1) for the target
population of lakes (<2000 ha) sampled in Subregion 3B
(Florida), Eastern Lake Survey-Phase I 40
4-5 F(x) and G(x) distributions of pH (closed system) for the target
population of lakes (<2000 ha) sampled in Region 1
(Northeast), Eastern Lake Survey-Phase I 41
4-6 F(x) and G(x) distributions of pH (closed system) for the
target population of lakes (<2000 ha) sampled in Region 2
(Upper Midwest), Eastern Lake Survey-Phase I 42
4-7 F(x) and G(x) distributions of pH (closed system) for the
target population of lakes (<2000 ha) sampled in Subregion 3A
(Southern Blue Ridge), Eastern Lake Survey-Phase I 43
xi
-------
Figures (continued)
Number Page
4-8 F(x) and G(x) distributions of pH (closed system) for the
target population of lakes (<2000 ha) sampled in Subregion 3B
(Florida), Eastern Lake Survey-Phase I 44
4-9 Comparisons among subregions of cumulative frequency
distributions [F(x>] for pH (closed system). Eastern Lake
Survey-Phase I 47
4-10 Comparisons among subregions of cumulative frequency
distributions [F(x)] for ANC (/ueq L~1), Eastern Lake Survey-
Phase I 48
4-11 Classes of ANC U/eq L"1) in lakes sampled in Region 1
(Northeast), Eastern Lake Survey-Phase I 51
4-12 Classes of ANC (/ueq L~1) in lakes sampled in Region 2
(Upper Midwest), Eastern Lake Survey-Phase I 53
4-13 Classes of ANC (/ueq L"1) in lakes sampled in Subregions 3A
(Southern Blue Ridge) and 3B (Florida), Eastern Lake Survey-
Phase I 54
4-14 Classes of pH in lakes sampled in Region 1 (Northeast),
Eastern Lake Survey-Phase 1 56
4-15 Classes of pH in lakes sampled in Region 2 (Upper Midwest),
Eastern Lake Survey-Phase 1 57
4-16 Classes of pH in lakes sampled in Subregions 3A (Southern
Blue Ridge) and 3B (Florida), Eastern Lake Survey-Phase I 58
5-1 Closed system pH versus initial open system pH for
probability sample lakes and special interest lakes in all
regions, Eastern Lake Survey-Phase I 68
5-2 pH (closed system) versus ANC (/yeq L~1) for Region 1,
Eastern Lake Survey-Phase 1 69
5-3 pH (closed system) versus ANC (/ueq L"1) for Region 2,
Eastern Lake Survey-Phase 1 70
5-4 pH (closed system) versus ANC (/ueq L~1) for Subregion 3A,
Eastern Lake Survey-Phase 1 71
5-5 pH (closed system) versus ANC (/ueq L~1) for Subregion 3B,
Eastern Lake Survey-Phase 1 72
5-6 pH (air-equilibrated) versus ANC (/ueq L~1) for Region 1,
Eastern Lake Survey-Phase I 73
5-7 pH (closed system) versus sulfate (/ueq L"1) for probability
sample lakes in Subregion 1 A, Eastern Lake Survey-Phase I 76
5-8 pH (closed system) versus sulfate (/ueq L~1) for probability
sample lakes in Subregion 2B, Eastern Lake Survey-Phase I 77
5-9 pH (closed system) versus sulfate (/ueq L"1) for probability
sample lakes in Subregion 3B, Eastern Lake Survey-Phase I 78
5-10 Extractable Al (/ug L~1) versus pH (closed system) for Region 1,
Eastern Lake Survey-Phase 1 79
5-11 Extractable Al (;ug L"1) versus pH (closed system) for Region 2,
Eastern Lake Survey-Phase 1 80
XII
-------
Figures (continued)
Number Page
5-12 Extractable Al (//g L~1) versus pH (closed system) for
Subregion 3A, Eastern Lake Survey-Phase I 81
5-13 Extractable Al (//g L"1) versus pH (closed system) for
Subregion 3B, Eastern Lake Survey-Phase I 82
5-14 Ionic composition of selected lakes from Region 1,
Eastern Lake Survey-Phase 1 85
5-15 Ionic composition of selected lakes from Region 2
and Subregions 3A and 3B, Eastern Lake Survey-Phase 1 86
5-16 Lake water sodium concentration (//eq L'1) versus
lake distance (km) from coast for Region 1,
Eastern Lake Survey-Phase 1 88
5-17 Lake water sodium concentration (//eq L~1) versus
lake distance (km) from coast for Subregion 3B,
Eastern Lake Survey-Phase 1 89
5-18 Relationship of sodium (//eq L"1) to chloride (//eq L"1)
for lakes in Subregions 1D (A) and 3B (B),
Eastern Lake Survey-Phase 1 90
5-19 Relationship of calcium (//eq L"1) to magnesium (//eq L~1)
for lakes in Subregions 2C (A) and 3B (B),
Eastern Lake Survey-Phase I 92
5-20 Trilinear plots of major anions and cations in
Subregions 1 A, 1 B, 1C and 1 D,
Eastern Lake Survey-Phase 1 93
5-21 Trilinear plots of major anions and cations in
Subregions 1E, 2A, 28 and 2C,
Eastern Lake Survey-Phase 1 94
5-22 Trilinear plots of major anions and cations in
Subregions 2D, 3A, and 3B,
Eastern Lake Survey-Phase 1 95
5-23 Dissolved organic carbon (0-20 mg L"1) versus
color (0-200 PCU) for Subregion 1 E,
Eastern Lake Survey-Phase 1 96
5-24 Dissolved organic carbon (0-20 mg L"1) versus
color (0-200 PCU) for Subregion 2A,
Eastern Lake Survey-Phase 1 97
5-25 Dissolved organic carbon (0-20 mg L~1) versus
color (0-200 PCU) for Subregion 3A,
Eastern Lake Survey-Phase 1 98
5-26 Relationship of the sum of anions (//eq L~1) to the
sum of cations (/t/eq L~1) for lakes in Subregion 1 A,
Eastern Lake Survey-Phase 1 100
5-27 Relationship of the sum of anions (//eq L"1) to the
sum of cations (//eq L~1) for lakes in Subregion 3A,
Eastern Lake Survey-Phase 1 101
5-28 Relationship of the sum of anions (//eq L"1) to the
sum of cations (//eq L~1) for lakes in Subregion 2A,
Eastern Lake Survey-Phase 1 102
xiii
-------
Figures (continued)
Number Page
5-29 Population estimates of lake numbers with ANC <200 #eq L~1
by hydrologic type for all regions. Eastern Lake
Survey-Phase I 105
5-30 Population estimates for median ftop bar) and Qi (bottom bar)
DOC by three classes of hydraulic residence time
(<0.5, 0.5 - 1.0, and >1 yr, respectively) for lakes in the
Northeast (Region 1}, Eastern Lake Survey-Phase I ,,,...,,..,... 107
6-1 Classes of ANC (//eq L~1) in five selected subpopulations of
lakes within Florida (3B), Eastern Lake Survey-Phase I 118
XIV
-------
Tables
Number Page
2-1 Assigned Numbers and Names for Regions and Subregions,
Eastern Lake Survey-Phase 1 9
2-2 Non-target and Not Visited Lakes,
Eastern Lake Survey-Phase 1 10
2-3 Principal Chemical and Physical Measurements,
Eastern Lake Survey-Phase 1 14
2-4 Aliquot Preparation and Preservation Requirements,
Eastern Lake Survey-Phase 1 19
2-5 Distribution of Lake Samples to Analytical Laboratories,
Eastern Lake Survey-Phase 1 20
3-1 Descriptions and Applications of Quality Control Samples,
Eastern Lake Survey-Phase 1 23
3-2 Descriptions and Applications of Quality Assurance Samples,
Eastern Lake Survey-Phase 1 28
3-3 Evaluation of Field Blank Data,
Eastern Lake Survey-Phase 1 29.
3-4 Estimated Within-Batch Precision from Field,
Trailer and Laboratory Duplicate Data,
Eastern Lake Survey-Phase 1 30
3-5 Among-Batch Precision Estimated from Field Natural Audit
Samples, Eastern Lake Survey-Phase I 31
4-1 Comparison of Sample Median pH and ANC to Estimated
Population Medians Using Weighting Factors,
Eastern Lake Survey-Phase 1 32
4-2 Comparison of Weights in Two Strata, 2A1 and 2A2,
Eastern Lake Survey-Phase 1 33
4-3 Data Quality Summary, Eastern Lake Survey-Phase I 34
4-4 Subregional and Regional Summaries of Estimated Target
Population Size (N~), Estimated Target Population Area (A),
and the Standard Errors (SE) of These Estimates,
Eastern Lake Survey-Phase 1 35
4-5 Description of Sample and Target Population
(Stratum Specific), Eastern Lake Survey-Phase I 36
4-6 Composition of the Alkalinity Map Classes, in Numbers
and Percentage of Lakes Having Measured ANC (//eg L~1)
in those Same Classes: 1) <100, 2) 100-200, 3) >200
for Lakes <2000 ha, Eastern Lake Survey-Phase I 49
4-7 Physical Lake Characteristics: Medians (M) and First
and Fourth Quintiles (Qi and Q.*), Eastern Lake
Survey-Phase I 50
XV
-------
Tables (continued)
Number Page
4-8 Population Estimates of Lake Type: Number of Lakes,
Eastern Lake Survey-Phase 1 50
4-9 Population Estimates of Lakes with ANC <0 ueq L~1,
Eastern Lake Survey-Phase 1 52
4-10 Population Estimates of Lakes with ANC <50 //eq L~1,
Eastern Lake Survey-Phase 1 52
4-11 Population Estimates of Lakes with ANC <200 /ieq L~1,
Eastern Lake Survey-Phase 1 52
4-12 Estimates of Numbers of Lakes with ANC <0, <50 and
<200 //eq L~1 by State, Eastern Lake Survey-Phase I 55
4-13 Population Estimates of the Proportion of Lakes with
ANC <200 //eq L~1 for Six Subpopulations,
Eastern Lake Survey-Phase 1 55
4-14 Population Estimates of Lakes with pH <5.0,
Eastern Lake Survey-Phase 1 56
4-15 Population Estimates of Lakes with pH <6.0,
Eastern Lake Survey-Phase 1 57
4-16 Estimates of Numbers of Lakes with pH <5.0 and <6.0
by State, Eastern Lake Survey-Phase I 59
4-17 Population Estimates of Lakes with Sulfate >50 //eq L"\
Eastern Lake Survey-Phase 1 59
4-18 Population Estimates of Lakes with Sulfate >150 //eq L"1,
Eastern Lake Survey-Phase 1 59
4-19 Population Estimates of Lakes with Calcium <50 //eq L"1,
Eastern Lake Survey-Phase 1 59
4-20 Population Estimates of Clearwater Lakes with Extractable
Aluminum >50 //g L"1, Eastern Lake Survey-Phase I 60
4-21 Population Estimates of Clearwater Lakes with Extractable
Aluminum >100 //g L~1, Eastern Lake Survey-Phase I 60
4-22 Population Estimates of Clearwater Lakes with Extractable
Aluminum >150 //g L"1, Eastern Lake Survey-Phase I 61
4-23 Population Estimates of Lakes with DOC <2 mg L"1,
Eastern Lake Survey-Phase 1 61
4-24 Population Estimates of Lakes with DOC >6 mg L"1,
Eastern Lake Survey-Phase 1 61
4-25 Primary Variables: First Quintiles (Qi), Medians (M),
and Fourth Quintiles (Q4), Eastern Lake Survey-Phase I 62
4-26 Secondary Variables (Nitrate, Ammonium and Total
Phosphorus): First Quintiles (Qi), Medians (M), and
Fourth Quintiles (Q4, Eastern Lake Survey-Phase I 62
4-27 Secondary Variables (True Color, Turbidity and Secchi Disk
Transparency): First Quintiles (Qi), Medians (M), and
Fourth Quintiles (Q4), Eastern Lake Survey-Phase I 63
XVI
-------
Tables (continued)
Number Page
4-28 Secondary Variables (Sodium, Potassium and Magnesium):
First Quintiles (Qi), Medians (M), and Fourth
Quintiles (Q4), Eastern Lake Survey-Phase I 64
4-29 Secondary Variables (Iron, Manganese and Total
Aluminum): First Quintiles (Qi), Medians (M), and
Fourth Quintiles (Q«), Eastern Lake Survey-Phase I 64
4-30 Secondary Variables (Silica, Dissolved Inorganic Carbon,
Chloride, Conductance and Bicarbonate): First
Quintiles (Qi), Medians (M), and Fourth Quintiles (Q.*),
Eastern Lake Survey-Phase 1 65
4-31 Sample Statistics for Special Interest Lakes by Region:
Minima (MIN), Medians (MED) and Maxima (MAX),
Eastern Lake Survey-Phase 1 66
5-1 Comparison of pH Measurements for Regular and Special
Interest Lakes, Eastern Lake Survey-Phase I 67
5-2 Eastern Lake Survey-Phase I (ELS-I) Population Estimates
for Selected Literature Definitions of Sensitivity 74
5-3 Regression Statistics for Log [Extractable Al]
(Molar, Dependent) versus pH (Independent) for
All Regions, Eastern Lake Survey-Phase I 77
5-4 Regression Statistics for ANC (Dependent) versus
Base Cations (Independent) by Subregion,
Eastern Lake Survey-Phase 1 83
5-5 Order of Major Cations and Anions Based on Population
Estimates of Concentrations at the 20th Percentile (Qi)
and Median Values, Eastern Lake Survey-Phase 1 87
5-6 Order of Major Anions by Subregion Based on the 20th
Percentile (Qi) and Median Concentrations Including
A~ as Unmeasured Anions, Eastern Lake Survey-Phase I 87
5-7 Regression Statistics for Sodium (Dependent) versus
Chloride (Independent) by Region for Concentrations
from 0 to 1000 //eq L"1, Eastern Lake Survey-Phase I 91
5-8 Regression Statistics for Calcium (Dependent) versus '
Magnesium (Independent) by Region for Concentrations
from 0 to 1000 //eq L"1, Eastern Lake Survey-Phase I 91
5-9 Regression Statistics for DOC (0-20 mg L"1, Dependent)
versus Color (0-200 PCU, Independent) by Subregion,
Eastern Lake Survey-Phase 1 99
5-10 Population Estimates of Qi (20th Percentile), Median and
Q4 (80th Percentile) Anion Deficit (//eq L"1) by
Subregion, Eastern Lake Survey-Phase I 99
5-11 Regression Statistics for the Sum of Anions (Dependent)
versus the Sum of Cations (Independent) by Subregion,
Eastern Lake Survey-Phase I 99
XVII
-------
Tables (continued)
Number Page
5-12 Regression Statistics for Anion Deficit (0-200 //eq L~1,
Dependent) versus DOC (0-20 mg L~1, Independent)
Computed without Metals (Al+3, Fe+3) and with Metals,
Eastern Lake Survey-Phase 1 103
5-13 Population Estimates of Lake Numbers by Hydrologic Types
with Associated Median Values of Chemical Parameters,
Eastern Lake Survey-Phase 1 104
5-14 Estimated Hydraulic Residence Time for Drainage Lakes
and Reservoirs by Subregion (Excludes Closed and
Seepage Lakes), Eastern Lake Survey-Phase I 106
5-15 Population Estimates Based on Selected Characteristics
and Their Associated Hydraulic Residence Time (RT),
Eastern Lake Survey-Phase 1 106
5-16 Population Estimates of Qi (20th Percentile), Median,
and Q4 (80th Percentile) of Selected Variables for
Acidic (ANC <0 /ueq L~1) Lakes for Regions 1 and 2,
and Subregion 3B, Eastern Lake Survey-Phase I 108
6-1 Medians and Interquintile Differences (CU - Qi = Qa) for
pH, ANC, Calcium, Sulfate and DOC, Eastern Lake
Survey-Phase I 111
XVIII
-------
Related Documents^
Anonymous. National Surface Water Survey, National Lake Survey-Phase I.
Research Plan. U.S. Environmental Protection Agency, Washington, DC,
(internal document), 1984.
Best, M. D., L. W. Creelman, S. K. Drouse, and D. J. Chaloud. 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, 1986.
Drouse, S. K., D. C. J. Hillman, L. W. Creelman, J. F. Potter, and S. J. Simon.
National Surface Water Survey, Eastern Lake Survey-Phase I. Quality Assurance
Plan. EPA/600/4-86/008, U.S. Environmental Protection Agency, Las Vegas,
Nevada, 1986.
Eilers, J. M., D. J. Blick, Jr., and M. D. DeHaan. 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, 1986.
Hillman, D. C. J., J. F. Potter, and S. J. Simon. 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, 1986.
Kanciruk, P., R. A. McCord, L. A, Hook, and M. J. Gentry. National Surface Water
Survey, Eastern Lake Survey-Phase I. Data Base Dictionary. Oak Ridge National
Laboratory, Technical Manual ORNL, 1986.
Morris, F. A., D. V. Peck, M. B. Bonoff, and K. J. Cabbie. 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, 1986.
Overton, W. S. National Surface Water Survey, Eastern Lake Survey-Phase I.
Data Analysis Plan. U.S. Environmental Protection Agency, Corvallis, Oregon,
1986.
1 Many of the documents are in draft form at the time of this publication.
xix
-------
Contributors
The Eastern Lake Survey-Phase I and this document represent the efforts of
many individuals. The primary contributors to this report are noted below.
Section 1: Introduction
P. E. Kellar, Radian Corporation
B. B. Emmel, Radian Corporation
Section 2: Methods
Design
W. S. Overton, Oregon State University
R. A. Linthurst, USEPA-EMSL, Research Triangle Park
Analytical Methods
D. V. Peck, Lockheed-EMSCO, Inc.
M. D. Best, Lockheed-EMSCO, Inc.
E. P. Meier, USEPA-EMSL, Las Vegas
Field Methods
D. V. Peck, Lockheed-EMSCO, Inc.
J. R. Baker, Lockheed-EMSCO, Inc.
W. E. Fallen, Battelle Pacific Northwest Laboratories
M. B. Bonoff, Lockheed-EMSCO, Inc.
R. E. Crowe, USEPA-EMSL, Las Vegas
Data Base Management
P. Kanciruk, Oak Ridge National Laboratory
SectionS: Quality Assurance
M. D. Best, Lockheed-EMSCO, Inc.
J. M. Eilers, Northrop Services, Inc.
P. Kanciruk, Oak Ridge National Laboratory
L. W. Creelman, Lockheed-EMSCO, Inc.
E. P. Meier, USEPA-EMSL, Las Vegas
Section 4: Results of Population Estimates
W. S. Overton, Oregon State University
D. F. Brakke, Western Washington University
D. H. Landers, USEPA-ERL, Corvallis
J. M. Eilers, Northrop Services, Inc.
S. A. Teague, Northrop Services, Inc.
Section 5: Results and Discussion of Associations among Variables
J. M. Eilers, Northrop Services, Inc.
D. F. Brakke, Western Washington University
D. H. Landers, USEPA-ERL, Corvallis
XX
-------
Section 6: Regional and Subregional Characteristics
D. F. Brakke, Western Washington University
J. M. Eilers, Northrop Services, Inc.
D. H. Landers, USEPA-ERL, Corvallis
Section 7: Summary Observations
D. F. Brakke, Western Washington University
J. M. Eilers, Northrop Services, Inc.
D. H. Landers, USEPA-ERL, Corvallis
XXI
-------
A ckno wledgments
The successful completion of so large and complex a field project as the Eastern
Lake Survey would have been impossible without the dedicated efforts of many
people. The short time between conceptualizing and implementing the Survey,
its scale, and its relevance in directing future research efforts and policy
decisions necessitated the involvement and dedication of many talented people.
The authors can acknowledge only a few of the contributors who played key roles
in the Survey's design and execution. For those who participated but are not
mentioned by name, your efforts and dedication are likewise acknowledged.
We especially thank Richard T. Dewling (State of New Jersey), formerly Deputy
Regional Administrator of EPA Region II. He provided direction during design and
implementation. His critical insight led repeatedly to valuable and timely
corrective actions. Without his presence and guidance, the Survey could not
have been accomplished so successfully.
The personal interest of Courtney Riordan, Director of the Office of Acid
Deposition, Environmental Monitoring and Quality Assurance, EPA Office of
Reseach and Development (ORD) was especially important to the Eastern Lake
Survey.
William Ruckleshaus and Alvin Aim, then Administrator and Assistant
Administrator of EPA, respectively, provided continual support and leadership
for the Eastern Lake Survey. Gary Foley, Director of the Acid Deposition and
Atmospheric Research Division and Raymond Wilhour of EPA-ORD offered
valuable assistance throughout this project in many capacities. Josephine
Huang (EPA-ORD) helped in many ways, particularly with administrative
guidance.
David Bennett (EPA-ORD) and Richard Wright (Norwegian Institute for Water
Research, Oslo, Norway) provided insight and encouragement in the early
design phases.
Kenneth Stoller (EPA-Region II) expertly coordinated the complex network of
field activities. Assistance in sampling was directed by the following people who
served as Base Coordinators and deputies and collectively made the field
operations a success:
EPA Region
Staff
Region I (Boston, MA)
Region II (Edison, NJ)
Region III (Philadelphia, PA)
Region IV (Atlanta, GA)
Region V (Chicago, IL)
Region IX (San Francisco, CA)
Region X (Seattle, WA)
Raymond Thompson, Daniel Murray
Rollie Hemmitt, John Alonso,
Randy Braun
Walter Graham
Ronald Raschke, Lawrence Brannan,
Andrew Peake
Phillip Gehring, Gerry Golubski
Arnold Den, Kenneth Greenberg
Lee Marshall, David Tetta
XXII
-------
Helicopter support was arranged by the Office of Aircraft Services, U.S.
Department of the Interior. Wes Kinney (EPA-EMSL/Las Vegas) coordinated
training of sampling personnel. The U.S. Forest-Service assisted in providing
access to the Boundary Waters Canoe Area in Minnesota. Terry Haines of the
U.S. Fish and Wildlife Service gave assistance with preliminary field studies.
Many state agencies (New York Department of Environmental Conservation,
Maine Department of Environmental Protection, Wisconsin Department of
Natural Resources, and Florida Department of Environmental Regulation) and
universities (Oregon State University, Western Washington University, and the
State University of New York at Oswego, among others) participated in the
development of the Survey. The states of Maine and Wisconsin provided
additional assistance with field sampling. All are acknowledged for their input to
and support of the project.
Lockheed-EMSCO, Inc., Northrop Services, Inc., and Radian Corporation
provided important planning, logistical, quality assurance and technical support
to the ELS-I management personnel. At Lockheed-EMSCO: Kenneth Asbury,.
Kevin Cabbie, and Stephen Pierett were responsible for procurement, base site
establishment and logistics, respectively; Steven Simon, Lynn Creelman, and
Sevda Drouse guided the quality assurance program; Franklin Morris and Daniel
Hillman assisted in the selection and development of sampling and analytical
methodologies. Sharon Teague (Northrop) assisted in many aspects of data
analysis and report preparation. Robert Cusimano (Northrop) assisted with the
development of field sampling protocols and assisted in training. Andrew Kinney
(Northrop) is acknowledged for his dedication to the implementation of the lake
selection process. Among the staff at Radian: Janice Stafford and Meredith
Haley are especially thanked for their superior typing support; Barbara Emmel
and Penelope Kellar were instrumental in the organization, editing, and
production of this report and their dedication and tolerance are greatly
appreciated. The authors are particularly indebted to Penelope Keflar. Through-
out the preparation of the report, and especially during the final days of the report
preparation, her production coordination, editing, rewrites, and attention to
detail were invaluable in producing this report.
The contribution of Rene Hinds in editing the final report is gratefully
acknowledged.
Professional consultants (Tim Webb of Environmental Systems and Systems
Applications, Clayton Creager of Kilkelly Environmental Associates, Kent
Thornton of FTN and Associates, and Alison Pollack and Thomas Permutt of
Systems Applications, Inc.) assisted with the planning and implementation of
the project, quality assurance and data analysis. In particular, Kent Thornton and
Tim Webb are specifically acknowledged for assisting in the development of the
draft research plan that made this Survey possible.
Walter Liggett of the National Bureau of Standards gave valuable advice on
statistical issues. James Kramer (McMaster University), Al Lefohn (ASL
Associates), and Pierre Sprey are thanked for reviewing the quality assurance
program. Jeffrey White, Indiana University, assisted with data analysis for
aluminum. Sample analysis for methods comparisons was conducted by: the
Ontario Ministry of the Environment, Rexdale, Ontario, Canada; Canada Centre
for Inland Waters, Burlington, Ontario, Canada; and Arne Henriksen at the
Norwegian Institute for Water Research, Oslo, Norway.
Data management support was coordinated through the Environmental
Sciences Division of Oak Ridge National Laboratory (ORNL) with assistance from
Tricia Gregory (Science Applications International Corporation) and John
Fountain, David Hoff, Gene Wilde, and Carol Brown (Lockheed-EMSCO).
XXIII
-------
Edward S. Deevey, Jr., of the University of Florida, James Galloway of the
University of Virginia, Steven A. Eisenreich of the University of Minnesota, and
Steven A. Norton of the University of Maine, are recognized for the constructive
suggestions they made to improve the quality of the report. The authors extend
their appreciation to these four reviewers and to the many others who offered
valuable comments during the review process.
To all of these people the authors extend their sincerest appreciation and
gratitude.
XXIV
-------
Executive Summary
The Eastern Lake Survey-Phase I (ELS-I) was conducted in the fall of 1984 as a
part of the U.S. Environmental Protection Agency's (EPA's) National Surface
Water Survey (NSWS). The NSWS is a contribution to the National Acid
Precipitation Assessment Program, which is charged by the U.S. Congress to
provide policy makers with sound technical information regarding the effects of
acidic deposition.
The ELS-I had three primary objectives:
1. Determine the percentage (by number and area) and location of lakes that
are acidic in potentially sensitive regions of the eastern U.S.
2. Determine the percentage (by number and area) and location of lakes that
have low acid neutralizing capacity in potentially sensitive regions of the
eastern U.S.
3. Determine the chemical characteristics of lake populations in potentially
sensitive regions of the eastern U.S. and provide the data base for selecting
lakes for further study.
To accomplish these objectives, a water sample was collected from each of 1612
lakes. This subset of lakes was selected from within three regions of the eastern
U.S. (the Northeast, Upper Midwest, and Southeast) expected to exhibit low
buffering capacity. Each region was divided into subregions, shown below:
Region 1 :
Northeast
1A: Ad rondacks
1B: Poconos/Catskills
1 C: Central New
England
1D: Southern New
England
1E: Maine
Region 2:
Upper Midwest
2A: Northeastern
Minnesota
2B: Upper Peninsula
of Michigan
2C: Northcentral
Wisconsin
2D: Upper Great Lakes
Area
Region 3:
Southeast
3 A: Southern Blue Ridge
3B: Florida
Each subregion was further stratified by alkalinity map class, which differentiated
among areas within each subregion based on the surface water alkalinity range
expected to dominate in different areas within these subregions.
A suite of chemical variables and physical attributes thought to influence or be
influenced by surface water acidification was measured for each lake. The
results of these measurements form the ELS-I data base.
The ELS-I design, in which lakes were selected by a systematic random process
from the population of lakes in the regions investigated, permits the use of ELS-I
data base to estimate the chemical status of lakes within a specific region or
subregion. Additionally, the data base can be used to investigate correlative
relationships among chemical variables on a regional basis.
The report which follows, Characteristics of Lakes in the eastern United States,
consists of three volumes. Volume I, Population Descriptions and Physico-
XXV
-------
Chemical Relationships, provides details about ELS-I design and its implementa-
tion, presents data collected in the ELS-I, discusses results obtained and draws
conclusions about these results. Volumes II and III contain descriptive statistics
for each lake sampled and a data compendium of site characteristics and
chemical variables.
This report is not intended to be interpretive. Rather, its purpose is to describe
the results and to make the ELS-I data available to researchers and policy makers
as more in-depth analyses and interpretive efforts are being undertaken.
Analyses will be performed in subsequent phases of the EPA Aquatic Effects
Research Program and by independent researchers.
The use and interpretation of any data set are restricted by the design, the quality
of the data obtained and the sampling protocols, which are presented in detail in
Sections 2, 3, and 4. These aspects of the Survey should be well understood
before drawing conclusions both within and beyond the scope of the original
objectives. For example, these data alone may not be sufficient to determine
causality. However, Survey data, coupled with data from ongoing and future
projects, are expected to significantly advance our understanding of the
relationship between acidic deposition and lake water chemistry.
Selected Results
The first two summary observations presented below address the first two ELS-I
objectives. The remaining summary observations address the third objective of
the Survey. These observations lead to hypotheses that can be tested in
subsequent phases of the National Surface Water Survey and/or the Aquatic
Effects Research Program.
It should be noted that the numbers and percentages of lakes cited here are
population estimates.
Extent and Location of Acidic and Low pH Lakes
The subregions in the eastern U.S. that contain the largest proportion of acidic
(ANC <0 /ueq L~1) and low pH (<5.0) lakes are the Adirondaks (1 A), the Upper
Peninsula of Michigan (2B), and Florida (3B).
Acidic Lakes
• Within the Northeast (Region 1), the Adirondacks (1A) had the largest
estimated number (138) and percentage (11 %) of lakes with ANC <0/ueqL~1,
followed by Southern New England(1D; 5%), and the Poconos/Catskills(1B;
5%). Maine (1 E) had the lowest percentage of acidic lakes (<1%). Most acidic
lakes in the Adirondacks occurred in the western portion of the subregion.
• In the Upper Midwest (Region 2), 10 percent of the lakes in the Upper
Peninsula of Michigan (2B) had ANC <0 //eq L~1, and three percent in
Northcentral Wisconsin (2C) were acidic. In Northeastern Minnesota (2A)
and the Upper Great Lakes Area (2D) no acidic lakes were sampled.
• In the Southeast (Region 3), no acidic lakes were sampled in the Southern
Blue Ridge (3A). In contrast, an estimated 22 percent of the lakes in Florida
(3B) had an ANC <0 yueq L'1.
• Acidic lakes in the Northeast had higher concentrations of sulfate, calcium,
and extractable aluminum than did acidic lakes in the Upper Midwest and
Southeast.
Low pH Lakes
The estimated number of lakes and lake area with low pH (pH <5.0) also varied
substantially among and within regions.
xxvi
-------
• Within the Northeast, the Adirondacks (1A) had the largest estimated
number (128) and percentage (10%) of lakes with pH <5.0. Subregion 1D
(Southern New England) contained the second highest estimated number
(66) and percentage (5%) and the largest area (2295 ha, 6%) of low pH lakes.
Maine (1E) had the fewest lakes (8, <1 %) and the least lake area (95 ha) with
pH <5.0.
• I n the Upper Midwest, no lakes with pH<5.0 were observed in Northeastern
Minnesota (2A) or the Upper Great Lakes Area (2D). The Upper Peninsula of
M ichigan(2B) was estimated to contain 99 lakes with pH<5.0, representing
nearly the same proportion as in the Adirondacks (9% and 10%, respectively).
• In the Southeast, no lakes with pH <5.0 were sampled in the Southern Blue
Ridge (3A). Florida (3B) had the highest estimated number and percentage of
lakes (259, 12%) and the largest estimated lake area with pH <5.0.
Extent and Location of Low ANC Lakes
As observed with the estimates of low pH lakes, the estimated number of lakes
with low ANC varied among and within regions.
• Within the Northeast, the Adirondacks (1A) contained the highest per-
centages of lakes with ANC <50 //eq L"1 and <200 //eq L"1 (35% and 70%,
respectively). Central New England (1C) and Maine (1 E) contained the next
highest percentages of lakes among all ELS-I subregions with ANC <200
//eq l~1 (68% and 67%, respectively).
• Northcentral Wisconsin (2C) contained the highest percentage (41%) of
lakes with ANC <50//eq L"1 among all subregions. Northeastern Minnesota
(2A) and Northcentral Wisconsin contained the highest percentage of lakes
in the Upper Midwest with ANC <200 //eq L~1 (57%). Although the Upper
Great Lakes Area (20) contained the lowest percentages in the Upper
Midwest of lakes with ANC <200//eq L"1, it contained the largest number of
lakes among all ELS-I subregions in this category (1411).
• The Southern Blue Ridge (3A) contained the lowest percentage (1%) and
number (4) of lakes in the ELS-I with ANC <50 //eq L~1 and the lowest
number of lakes with ANC <200 //eq L"1 among all subregions. Florida (3B)
contained the highest number of lakes among all ELS-I subregions with ANC
<50 //eq L"1, and the second highest number of lakes with ANC <200 //eq
L'1.
Chemical Characterization
Sulfate
Sulfate concentrations in lakes were greatest in Florida and the southern
portions of the Northeast. No linear relationship between lakewater sulfate and
pH or ANC was evident in any region. High concentrations of sulfate were found
at low and high pH values.
• Sulfate concentrations were relatively high in the Northeast Region (median
concentration (M) = 115.4 //eq L"1). Within the Northeast, sulfate concentra-
tions were highest in the Poconos/Catskills (1B; M = 159.3 //eq L~1) and
Southern New England (1 D; M= 141.1 //eq L"1). The lowest sulfate values
were observed in Maine (1 E; M = 74.6 //eq L"1).
• The median sulfate concentration in the Upper Midwest Region was half
that of the Northeast. Median sulfate concentrations also varied among
subregions within the Upper Midwest, ranging from 50.1 //eq L'1 in the
Upper Great Lakes Area (2D) to 77.7 //eq L~1 in the Upper Peninsula of
Michigan (2B).
xxvii
-------
• In the Southeast, the Southern Blue Ridge Subregion (3A) contained few
lakes with high sulfate (22 or 8% with S04~2 ^150 /ueq L"1}. This subregion
also had the lowest median sulfate concentration, 31.8 (jeq L~1. Floride (3B)
contained the largest number of lakes with high sulfate concentrations(846
or 40% with SO*'2 >150 yueq L~1). Subregion 3B also had the most variable
sulfate concentrations of any subregion.
Calcium
Calcium concentrations were lowest in the Upper Midwest and Florida lakes.
• Within the Northeast Region, Southern New England (10) had the highest
percentage and number of lakes with calcium concentrations <50 /ueq L~1
(10%; 133). The Adirondacks (1 A) contained the second highest percentage
and number (8%; 108) of low calcium lakes (<50 /ueq L"1).
• Northcentral Wisconsin (1C) contained the highest percentage (22%) and
second highest number (34) of low calcium lakes among all ELS-I
subregions. The Upper Peninsula of Michigan (2B) contained the second
highest percentage (16%) of low calcium lakes and the Upper Great Lakes
Subregion (2D) contained the second highest number (256) of low calcium
lakes in the Upper Midwest.
• In the Southeast, 12 percent of the lakes in the Southern Blue Ridge(3A) had
low concentrations of calcium, whereas in Florida (3B), 19 percent of the
lakes were in this group. Florida contained the highest number (402) of low
calcium lakes among all subregions.
Extractable Aluminum
Extractable aluminum concentrations were higher in lakes with lower pH
values, and higher in the Northeast than in other regions.
• The largest estimated number of clearwater lakes (true color <30 PCU)
having extractable aluminum concentrations >150 /ug L"1 occurred in the
Adirondacks (1 A; 82 lakes or 10%). Few lakes in the Poconos/Catskills(1 B; 3
lakes or <1 %) and Southern New England (1D; 7 lakes or 1 %) had extractable
aluminum >150 fjg L'1. No clearwater lakes sampled in Maine (1E) had
extractable aluminum concentrations >50 jug L~1.
• Extractable aluminum concentrations in clearwater lakes were lower in the
Upper Midwest (80th percentile = 8.5 vg L~1) than in the Northeast (80th
percentile = 11.6 yug L~1). Extractable aluminum was lowest in clearwater
lakes in Northeastern Minnesota (2A; 80th percentile = 3.0 fjg L~1), and
highest in clearwater lakes in the Upper Penninsula of Michigan (2B; 80th
percentile = 11.9 fjg L"1).
• Extractable aluminum concentrations in clearwater lakes were low in the
Southern Blue Ridge (3A; 80th percentile = 2.5 fjg L"1). In Florida (3B),
clearwater acidic lakes had lower extractable aluminum concentrations
(80th percentile = 18.6 /ug L~1) than did clearwater lakes in the Adirondack
Subregion (1 A; 80th percentile = 29.4 /ug L"1).
• In each region extractable aluminum concentrations were higher at lower
pH values. The Northeast had the greatest increase in extractable aluminum
with decreasing pH and Florida the least increase at low pH values.
Dissolved Organic Carbon
Dissolved organic carbon (DOC) concentrations did not correlate with the
distribution of acidic or low ANC lakes.
xxviii
-------
• in the Northeast, as in other regions, 80 percent of acidic lakes contained
concentrations of DOC <5 mg L~1. A positive relationship existed between
pH and DOC. Those lakes with highest DOC concentrations were drainage
lakes with short hydraulic residence times and high ANC.
9 In the Upper Midwest, most acidic lakes, especially those in the Upper
Peninsula of Michigan (2B) and Northcentral Wisconsin (2C), were
clearwater, low DOC, seepage lakes. Lakes in Northeastern Minnesota (2A)
had the highest concentrations of DOC in the Upper Midwest and no acidic
lakes were sampled in this subregion.
• in the Southeast, only the lakes within the Okefenokee Swamp exhibited a
strong association between low pH and DOC. No apparent relationship
between pH and DOC was evident in Florida (3B) lakes.
Major Cations and Anions
The anions were most useful in characterizing differences in the relative
importance of major ions among regions and subregions.
• In the Northeast, sulfate was the predominant anion at the 20th percentile in
three of the subregions (Adirondacks; 1A; Poconos/Catskills, 1B; and
Central New England, 1C)V Sulfate was also the dominant anion at the
median value in the Adirondacks.
9 In Maine (1E), bicarbonate ion concentrations exceeded sulfate at both the
20th percentile and the median.
• Chloride was the dominant anion in Southern New England (1D) at both the
20th percentile and median values estimated for the population.
9 Bicarbonate was the dominant anion at the 20th percentile and median
values in the Upper Midwest, with the exception of the Upper Peninsula of
Michigan (2B) and Northcentral Wisconsin (2C), where sulfate was
dominant at the 20th percentile.
« The ionic composition of lakes in Florida (3B) was similar to that of lakes in
Southern New England (1 D) in that sodium was the dominant cation and
chloride the dominant anion at the 20th percentile. Total ionic concentration
of many Florida lakes was high.
• Organic anions, as indicated by anion deficit, were not the dominant anions
in any subregion at either the 20th or 50th percentiles. Concentrations of
organic anions were lowest in the Northeast.
Future Studies
Examination of the results of the ELS-I Survey presented in this report are largely
descriptive but lead one to forumulate hypotheses that can be tested with this
data base, singularly or combined with other data. The statistical design of the
Survey makes it possible to test hypotheses related to acidification using
regional data and relate the results to defined, regional lake populations. Five
statements follow which are the results of the ELS-I Survey. Beneath each
statement is a related question that should be addressed in the future:
• Sulfate concentrations in lakes across the Northeast and the Upper Midwest
show an apparently strong relationship with the general patterns of sulfate
deposition as measured by the National Trends Network.
What is the nature of the relationship between lake chemistry and
atmospheric depositon of sulfate?
xxix
-------
• The majority of acidic lakes in all three regions contained relatively low
concentrations of organic acids.
How important are the contributions of organic acids in explaining the
occurrence of acidic lakes?
• Some portions of the coastal areas of the Northeast contained moderate
numbers of acidic lakes.
To what degree can the acidity of these coastal lakes be attributed to a
neutral salt effect from sea spray deposition?
• The estimated hydraulic residence times for clearwater lakes were
approximately 3 times greater than for darkwater lakes. Residence time was
inversely related to DOC.
Does an apparent difference in hydrology between clearwater, acidic lakes
and darkwater, higher ANC lakes indicate that acidic lakes generally are not
derived from darkwater lakes?
• Florida (3B) contained the largest proportion of acidic lakes and the
chemistry of many Florida lakes differed considerably in many respects from
lakes in the Northeast, Upper Midwest and Southern Blue Ridge (3A).
To what degree are the acidic lakes in Florida affected by acidic deposit/on,
and are other factors important in explaining the occurence of acidic lakes in
Florida?
xxx
-------
Section 1
Introduction
The relationship between acidic deposition and the
acidification of surface waters has become an im-
portant environmental issue in the United States
and other nations. To assess the impact of acidic
deposition on aquatic resources, many individual
studies and analyses of historical data have been
performed. The question of whether the pH and
alkalinity or acid neutralizing capacity (ANC) of sur-
face waters have been affected by acidic inputs of
anthropogenic origin has been extensively re-
viewed, analyzed and debated (Drablos and Tollan
1980; NAS 1981; NRCC 1981; U.S./Canada 1983).
Evidence from site-specific studies in the United
States has been published that is consistent with
the hypothesis that reductions in the pH and ANC of
surface waters have occurred over time. In the
analyses of chemical trends in lakes of the Adiron-
dack Region of New York and in New England sev-
eral authors have concluded that pH or alkalinity, or
both, has declined historically from approximately
the 1930's to the 1970's (Schof ield 1976; Davis et al.
1978a; Pfeiffer and Festa 1980; Norton et al. 1981;
Haines and Akielaszek 1983).
Despite these attempts to quantify changes in
aquatic chemistry, the degree to which surface
waters in the United States are threatened or have
been affected by acidic deposition remains un-
known. Uncertainty in extrapolating the results of
specific studies to a regional or national scale con-
tinues, largely due to: the unknown degree to
which individual study sites represent larger re-
gional populations, unknown bias in lake selection,
absence of measurements of chemical variables
that are critical in assessing chemical or biological
effects, difficulty in comparing data collected by dif-
ferent and sometimes unknown methods, and in-
consistent documentation of quality assurance pro-
tocols.
Alkalinity and ANC have been used as indices of
surface water sensitivity to acidic deposition (Alt-
shuller and McBean 1980; Altshuller and Linthurst
1984). The actual sensitivity of a lake or stream to
acidification, however, depends on the ANC gener-
ated both within the lake and its watershed. Hence,
because many physical, chemical and biological
factors, both aquatic and terrestrial, collectively
determine the biotic composition and chemical
environment within lakes, the response of an aquatic
ecosystem to acidic depostion is a composite of many
factors. Single-factor indices of potential sensitivity
are, therefore, limited as indicators of response to
acidic inputs.
Recognizing the limitations associated with avail-
able data and the complexity of predicting surface
water response to acidic deposition, the U.S. Envi-
ronmental Protection Agency initiated the National
Surface Water Survey (NSWS) in 1983. The NSWS,
which includes both lakes and streams, has three
primary goals and three distinct phases: 1) quantify
the present chemical status of surface waters in the
U.S. (Phase I), 2) assess the temporal and spatial
variability in aquatic chemistry and define the key
biological resources associated with surface waters
(Phase II), and 3) identify temporal trends in surface
water chemistry and biology (Phase III). Each sub-
sequent phase builds on the results of the previous
one, ultimately identifying those lake or stream
populations upon which to base a regionally char-
acteristic, statistically sound long-term monitoring
project designed to study long-term trends in
chemistry and biological resources.
Phase I of the lake survey has two components, the
Eastern Lake Survey (ELS-I) conducted in the fall of
1984 and the Western Lake Survey conducted in the
fall of 1985. The results of the Western Lake Survey
will be published in a report scheduled to be com-
pleted in late 1986. Phase II of the Eastern Lake
Survey began in the spring of 1986. The existing EPA
lake sites for long-term monitoring are currently
being evaluated in light of the ELS-I data base for
possible inclusion in the future Eastern Lake Survey
(Phase III) Long-Term Monitoring Project.
The ELS-I was designed to provide the information
needed to assess the chemical status of lakes in
areas of the eastern U.S. containing the majority of
low alkalinity systems. Lakes were selected statisti-
cally from the population of lakes within the north-
eastern, southeastern and upper midwestern re-
gions of the U.S. Variables thought to influence or
be influenced by surface water acidification were
measured using standardized methods. The ELS-I
-------
data base may be used to investigate correlative
relationships among chemical variables on a re-
gional basis and to estimate the chemical status of
lakes within a specific region.
The primary objectives of the ELS-I were to:
(1) determine the percentage (by number and
area) and location of lakes that are acidic in
potentially sensitive regions of the eastern
U.S.;
(2) determine the percentage (by number and
area) and location of lakes that have low ANC in
potentially sensitive regions of the eastern
U.S.; and
(3) determine the chemical characteristics of lake
populations in potentially sensitive regions of
the eastern U.S. and provide the data base for
selecting lakes for future studies.
In addition to the primary objectives that guided the
approach, the ELS-I was desgined to alleviate un-
certainty in making regional scale assessments.
Several additional considerations guided the de-
sign:
(1) Because all lakes of interest could not be
sampled, a subset of lakes was statistically
selected from a population of lakes. Because .
small lakes«4 ha) could not be included in the
sampling frame for a large synoptic survey,
they were not included in the population
considered.
(2) Sampling and analytical methods were uni-
form to ensure comparability of data.
(3) Because pH and ANC are ony two of the
chemical variables that may be important in
understanding lake sensitivity to acidic depo-
sition, many additional variables were meas-
ured.
(4) Measurements had to be of known quality;
population estimates had to be of known cer-
tainty. Therefore, the statistical and analytical
methods were well defined and carefully im-
plemented.
/
(5) Because regional, rather than individual, lake
characteristics were the primary focus of the
Survey, the ability to make comparisons
among lakes rather than within lakes was em-
phasized in the design.
(6) The ELS-I was designed to provide a data
base and establish a statistical basis for se-
lecting subsets of lakes for more detailed
study in future phases.
The purpose of this report is to present results ob-
tained from the ELS-I. Results will be used to frame
subsequent surveys designed to quantify biological
resources and temporal variability in chemistry,
and to implement a long-term monitoring program.
The report is organized as follows: Section 2, Meth-
ods; Section 3, Quality Assurance; Section 4, Re-
sults of Population Estimates; Section 5, Results
and Discussion of Associations among Variables;
Section 6, Regional and Subregional Characteris-
tics; Section 7, Summary Observations; Section 8,
References; and Section 9, Glossary. In addition,
reports presenting more detailed information on
methods are referenced. Two companion volumes.
Volume II: Lakes Sampled and Descriptive Statis-
tics for Physical and Chemical Variables and Vol-
ume III: Data Compendium of Site Characteristics
and Chemical Variables, provide the data used to
compile this report.
-------
Section 2
Methods
2.1 Design
2.1.1 Data Quality Objectives
Data quality objectives were developed to guide the
selection of sampling and analytical procedures for
the Eastern Lake Survey - Phase I (ELS-I). These
objectives defined goals for precision and bias both
for the procedures used in making population esti-
mates and for the methods used in analyzing chem-
ical variables. The data quality objectives guided
the statistical design of the Survey, the selection of
analytical methods and the design of the quality
assurance program (Orouse et al. 1986).
2.1.2 Lake Representation
A critical issue in the design of the ELS-I was the
representation of a selected lake. If a single water
sample can adequately represent the chemistry of a
lake to satisfy the specific objectives of a study, a
large number of lakes can be sampled. If multiple
water samples are needed on a single occasion,
then a reduction in the number of sample lakes
must be considered. If multiple occasions are
needed to represent the chemistry of a single lake,
the number of sample lakes must be reduced pro-
portionally.
It is obvious that one sample, from one location, at
one time of the day, in a specific season of a partic-
ular year, cannot characterize the complex chemi-
cal dynamics of a lake. Such a sample is justified
only in the sense that it is an index to the essential
characteristics of the lake. But even if two samples
are taken, or three, they remain only indices be-
cause understanding the dynamics of a single lake
requires far more detailed study. This study was
designed to describe populations of lakes. There-
fore, each lake must be represented in that popula-
tion description in a manner that captures its
essence, but such that the number of lakes that can
be sampled is maximized. The single index sample
maximizing both lake number and spatial coverage
on a large geographic scale was therefore deemed
the most appropriate choice for addressing the col-
lective objectives of the ELS-I.
To enhance the utility of the index sample, careful
consideration was given to location and season.
The sampling window was designated as the fall
season, just after turnover. Spatial variation within
the lake is reduced at this time. Sampling at the
apparently deepest part of the lake was intended to
provide a sample from the dominant water mass.
Therefore, the combination of a fall season sampling
period and collecting a sample near the lake center at
the apparently deepest part, appeared to be the best
protocol to provide the needed sampling character-
istics.
The perspective that each lake is represented by an
index chemistry, rather than, for example, mean
chemistry or some other integration over time and
space, is important in interpreting the results pre-
sented in this report. The population descriptions
represent and characterize the chemistry of a popu-
lation of lakes, as though every lake in the popula-
tion had been sampled in the same manner as the
sampled lakes. Thus the resulting frequency and
areal distributions for the chemical parameters
(Sections 4.2 and 4.3) represent an index to water
mass chemistry for the population of lakes that can
be interpreted only through study of the predictive
capacity of that index.
Studies of temporal and spatial pattern in a sub-
sample of lakes (Phase II) will provide an opportu-
nity to assess the soundness of these decisions and
to evaluate more fully the information that may be
gained using the index concept.
2.1.3 Statistical Design
The sampling plan for ELS-I employed a stratified
design, with equal allocation of number of sample
lakes to strata. Three strata were sampled with
greater intensity to obtain increased precision (Sec-
tion 2.2.3). This design is consistent with the objec-
tive of describing the populations of lakes within
the individual strata with roughly the same preci-
.sion. Lakes were selected from each stratum by
systematic sampling of an ordered list (Section
2.2.2) following a random start.
The choice of sample size, 50 or more target lakes
per stratum (Section 2.2.3), was based on the judg-
ment that this sample size was expected to yield
adequate precision for the 33 strata in the ELS-I. To
evaluate expected precision from various sample
sizes, the key statistic was chosen to be pc, the esti-
-------
mated proportion of lakes with a value of the vari-
able X below some reference value, Xc (for exam-
ple, the percentage of lakes with pH <6.0). Within a
single stratum, pc can be estimated by n^n, where
nc is the number of lakes in the sample that meet
that condition, and n is the sample size. Within a
single stratum, nc can be treated as a binomial vari-
able. The variance of the binomial variable is great-
est when p = 0.5. Assessment of precision at this
value yields a worst case condition. The results of
the comparative assessment are summarized be-
low, using the normal approximation and a one-
sided confidence limit:
Maximum
Standard Error (p)
Maximum Upper
Confidence Limit
100
60
50
40
30
0.0500
0.0645
0.0707
0.0791
0.0913
P^
P^
P^
PH
PH
h 0.082
h 0.106
h 0.116
h 0.130
h 0.150
where: Maximum Standard Error (p) =
Based on this comparison, the population estimate
for the number of lakes within a stratum [F(x) curve
(Section 4.3.2.1)] was estimated to be within 12 per-
cent for a sample size of 50. A number of other
considerations are involved in actually setting the
confidence limits; the above assessment served
only as a planning device. It is also of interest that
at the sample size of 50, the least upper confidence
bound occurs when nc (the number of lakes meet-
ing the condition defined by Xc) is zero, and is
within 6 percentage points (0.06).
2.1.4 Identification of Study Area
The population of lakes to be sampled was defined
as lakes located within those regions expected to
contain the most lakes in the U.S. characterized by
alkalinity <400 jteq L~1 (i.e., those areas where
acidic deposition would potentially have the most
effect). For the ELS-1, the boundaries of three re-
gions, judged to contain about 95 percent of the
lakes of lower alkalinity in the eastern U.S., were
delineated using a national map of surface water
alkalinity (Omernik and Powers 1983). The North-
east, the Upper Midwest, and the Southeast, as de-
fined in Figure 2-1, were selected as the study areas
in which to conduct the ELS-I.
2.2 Lake Selection
2.2.? Delineation of Strata
The three regions shown in Figure 2-1 represent the
first level of stratification in the design. Using re-
gion as a stratification factor ensured that each re-
gion containing low alkalinity lakes would be ade-
quately represented in the sample.
The second stratification factor, subregion, identi-
fied areas within each region that were expected to
be relatively homogeneous with respect to water
quality, physiography, vegetation, climate, and
soils. Based on geographic homogeneity, five sub-
regions (1A-1E) were identified in the Northeast,
four (2A-2D) in the Upper Midwest, and two (3A and
3B) in the Southeast (Figures 2-2 to 2-4). Descriptive
names of subregions were assigned (Table 2-1).
Subregion was used as a stratification factor to en-
sure that a representative sample was drawn from
each geographically distinct portion of a region.
The third stratification factor, alkalinity map class,
differentiated among areas within each subregion
based on the range of surface water alkalinity val-
ues expected to dominate in different areas. The
alkalinity map classes chosen were <100 n-eq L~1
(class 1), 100-200 |xeq L~1 (class 2), and >200 ixeq
L~1 (class 3). Spatial representations of the three
alkalinity classes within each region were derived
from preliminary versions of regional surface water
alkalinity maps (Omernik and Kinney 1985;
Omernik and Griffith 1985; Omernik 1985). An alka-
linity of 200 ^eq L~1 has been used (Galloway 1984)
as the boundary distinguishing between those sur-
face waters considered to be potentially "sensitive"
and those considered "insensitive" to long-term
acidification as a result of current levels of acidic
deposition. The choice of 100 fieq L~1 was based on
evidence that biological effects of acidification
might become apparent in the alkalinity range 10-
90 jieq L~1 (Haines and Akielaszek 1983).
In summary, each stratum is an alkalinity map class
within a subregion within a region. All three alkalin-
ity map classes were found within each of the
eleven subregions, so that a total of 33 strata (15 in
the Northeast, 12 in the Upper Midwest, and 6 in the
Southeast) were defined. The strata are coded by
region, subregion, and alkalinity map class. For ex-
ample, 1A2 designates the Northeast Region, the
Adirondack Subregion, and alkalinity map class 2.
2.2.2 Development of the Map Population and
Frame
Region, subregion, and alkalinity map class
boundaries were delineated and labeled on
1:250,000-scale U.S. Geological Survey (USGS) to-
pographic maps. The lake type, elevation, size, and
watershed size were evaluated as potential order-
ing factors by means of a mapping exercise to
check for spatial patterns. The selection and use of
these ordering factors as they apply to the genera-
tion of the frame (see below) are described in detail
in Omernik et al. (1986).
Each lake represented on the maps was then as-
signed a unique number. Lakes were numbered
consecutively within a mapping unit starting in its
4
-------
northwest corner and working east within a strip
one to four inches wide, depending on the density
of lakes represented on the maps. This process was
repeated from north to south until all lakes in the
mapping unit were numbered. Lakes in the next
mapping unit were then numbered, beginning with
the next consecutive number, until the entire stra-
tum was completed. In strata where mapping units
were not used, the process was the same as if the
entire stratum were one mapping unit. The final
number in each stratum was the total number of
lakes in the map population for that stratum.
The map population (frame population) represents
the universe of lakes considered for study in the
ELS-I. All population estimates computed in this
study refer to the map population (Section 2.2.4)
and do not represent conditions in lakes outside the
area of coverage or in systems not depicted on the
USGS maps used.
Figure 2-1. Regions and subregions surveyed during the Eastern Lake Survey-Phase I.
Upper Midwest
X
\
- A.
i
i
-j
i
i
*s
i
i.
)
vv
'•«
,'
i
\
>
Vi\
"N
/
<
-\
\
f
i.....
Subregion Boundary
-------
Figure 2-2. Northeastern subregions and alkalinity map classes. Eastern Lake Survey-Phase I.
Alkalinity Map Classes
(A*q L'1)
[Tj 200
— Subregion Boundary
*
Adirondacks{1 A) m
.-> ^i
Southern New England (10)
2.2.3 Probability Sample
Within each stratum a systematic random sample
was used to select lakes. Lake numbers were en-
tered into a computer file in numerical order as
labeled on the maps. In the strata in Subregions 1 A,
1C, 1E, and 2A, lakes were in spatial order within
ordering factor classes. In the remaining subre-
gions, lakes were simply in spatial order within
strata.
For each stratum the map population size was di-
vided by the prescribed sample size to obtain a
number (k). The first sample lake was then selected
at random between lakes 1 (one) and k. Thereafter,
every kth lake was selected. This procedure ensured
that each lake in the frame list of a stratum had an
equal probability of inclusion.
The prescribed sample size was 50 for all strata
except stratum 1A1 within the Adirondacks (60
lakes), stratum 1C1 within Central New England (70
lakes) and stratum 1E1 within Maine (90 lakes). Only
19 lakes occurred within alkalinity map class 1 in the
Southern Blue Ridge (stratum 3A1), so all were
included.
2.2.4 Identification and Refinement of Frame
Population
The map population was refined through the elimi-
nation of "non-target" lakes. The lakes in the prob-
ability sample were examined on 1:24,000-,
1:25,000- or 1:62,500-scale USGS topographic
maps. Categories of lakes that were collectively
termed non-target lakes included:
-------
(1) No lake present: lakes initially identified on
1:250,000-scale maps that did not appear on
more detailed, larger-scale maps.
(2) Flowing water: sites identified as lakes on
1:250,000-scale maps that appeared as points
on a stream on larger-scale maps. However, if
the small-scale maps were more recent than
the large-scale maps and the lake in question
was known to be a new reservoir, it was not
eliminated.
(3) Bay/Estuary (High conductance): lakes identi-
fied on 1:250,000-scale maps that appeared
as ocean embayments or estuaries on larger-
scale maps.
(4) Urban/lndustrial/Agricultural: lakes sur-
rounded by or adjacent to intense urban, in-
dustrial, or agricultural land use including
tailing ponds, water treatment lagoons, fish
hatcheries, and cranberry bogs.
(5) Marsh/Swamp: lakes identified on 1:250,000-
scale maps that appeared as swamps or
marshes on larger-scale maps.
(6) Too small (<4 ha): lakes identified on
1:250,000-scale maps that were less than ap-
proximately four hectares. Because the reso-
lution of most 1:250,000-scale maps was
about four hectares, this limit was estab-
lished for consistency. Lakes less than four
hectares are not represented by the popula-
tion descriptions.
After eliminating non-target lakes by examining
maps, additional lakes were selected, when neces-
sary, by applying the same systematic random
sampling process to the remaining lakes in the ini-
tial list frame. More than the prescribed number of
lakes were selected, since it was also anticipated
that additional lakes would be eliminated as non-
target lakes during field operations. Because the
Figure 2-3. Upper midwestern subregions and alkalinity map classes. Eastern Lake Survey-Phase I.
Alkalinity Map Classes
(//eq L'1)
m200
Subregion Boundary
^^^r^L^x? (fl*
^-^2^^^ C~f 3
Northeastern Minnesota (2A)
.,2
Upper Peninsula of Michigan (2B)
Northcentral Wisconsin (2C)
Upper Great Lakes Area (2D)
-------
Figure 2-4. Southeastern subregions and alkalinity map classes. Eastern Lake Survey-Phase I.
Alkalinity Map Classes
(/ueq L"1)
JLJ<100
F| 100-200
3~|>200
Subregion Boundary
-------
Table 2-1. Assigned Numbers and Names for Regions and Subregions, Eastern Lake Survey - Phase I
Region 1: Northeast Region 2: Upper Midwest Region 3: Southeast
1A: Adirondacks
1B: Poconos/Catskills
1C: Central New England
1D: Southern New England
1E: Maine
2A: Northeastern Minnesota
2B: Upper Peninsula of Michigan
2C: Northcentral Wisconsin
2D: Upper Great Lakes Area
3A: Southern Blue Ridge
3B: Florida
number of non-target lakes varied from stratum to
stratum, the final number of lakes selected within
strata varied.
The remaining lakes were provisionally designated
as target lakes with this designation to be further
refined as a result of information obtained during
field sampling. The categories of non-target lakes
eliminated by the field crews included all categories
above except the "too small" class. These were
defined as:
(1) No lake present: lakes visited that were found
to be dry.
(2) Flowing water: sites visited and found to be
streams.
(3) High conductance: lakes that, upon visitation,
were found to have measured specific con-
ductance >1500 jiS cm""1.
(4) Urban/lndustrial/Agricultural: lakes sur-
rounded by or adjacent to intense anthropo-
genic activities.
(5) Too shallow2: lakes that were too shallow to
obtain a clean (i.e., free of debris or sediment)
sample.
(6) Too small: this category was not used by field
crews (see above categories).
(7) Other: lakes that were inaccessible due to a
permanent feature of the lake that prevented
helicopters from landing safely (e.g., power
lines).
Some other lakes in the provisional field sample
were not visited, including those that were inacces-
sible due to bad weather, which were frozen, or to
which access permission was denied. These lakes
represent incompleteness in the sample and were
classified as "not visited" (could not be sampled)
only if the reasons for not sampling were unrelated
to a permanent feature of the lake. If a lake could
not be visited because of a permanent feature such
20riginal sampling protocols were designed to take samples only from
1.5 m. However, this protocol was modified in the field to allow samples
to be taken from 0.5 m in lakes where a 1.5-m clean sample could not be
taken, provided a 0.5-m clean sample could be taken (Section 4.2.2).
as shallowness, it was classified as non-target.
Those lakes not visited because they were frozen,
for example, did not warrant non-target classifica-
tion since this is not a permanent feature of the
lake. For statistical analyses, it was assumed that
lakes not visited had the same proportion of non-
target lakes as those found in the visited lakes.
Table 2-2 provides numbers of non-target lakes
identified during map examinations and during
field sampling, and reasons for their exclusion from
the target population. Categories and numbers of
lakes classified as not visited are also shown. These
definitions collectively identify the populations of
lakes about which conclusions can be drawn; they
further restrict the interpretation of the results as
does the identity of the map/frame population from
which the probability sample was drawn.
2.2.5 Special Interest Lakes
A number of lakes other than those chosen in the
probability sample were sampled during the ELS-I.
All lakes in the current EPA Long-Term Monitoring
Program, which is also part of the National Acid
Precipitation Assessment Program, were included
as special interest lakes. Others were included on
the basis of recommendations from the Acid Depo-
sition Trends Committee of the National Research
Council and from state and federal agencies.
Because these special interest lakes were not cho-
sen by the steps discussed in Section 2.2.3, the data
from these lakes were not included in population
estimates (Section 2.3.2). They were, however, in-
cluded in the analyses of associations among vari-
ables because such analyses are not prohibited in
the application of the design (Section 2.3.4). To dif-
ferentiate special interest lakes from those in the
probability sample, the latter are referred to as
"regular" lakes.
2.2.5 Final Lake List and Maps
Each selected lake was given a unique identification
(ID) number coded by the three levels of stratifica-
tion. For example, 1A2-034 designated the 34th lake
in Subregion A, in Region 1, alkalinity map class 2.
Lake ID numbers, lake names, geographical coordi-
nates, and map names were entered into computer
-------
Table 2-2. Non-Target and Not Visited Lakes, Eastern Lake Survey - Phase I
A. Non-target regular lakes determined from large-scale map examination
CATEGORIES 1A 1B 1C 1D 1E 2A 2B
2C
2D
3A
3B
TOTAL
NO LAKE PRESENT
FLOWING WATER
BAY/ESTUARY
URBAN/INDUSTRIAL
MARSH/SWAMP
TOO SMALL (<4 ha)
TOTAL
B. Non-target regular lakes
CATEGORIES
NO LAKE PRESENT
FLOWING WATER
HIGH CONDUCTANCE
URBAN/INDUSTRIAL
TOO SHALLOW
OTHER
TOTAL
1
8
0
0
5
23
37
determined
1A
2
0
0
2
7
1
12
C. Reasons that regular lakes were not
CATEGORIES
NO ACCESS PERMISSION
TIME/DISTANCE
CONSTRAINTS
BAD WEATHER
WRONG LAKE
HIGH TURBIDITY
FROZEN
TOTAL
1A
1
3
0
0
0
0
4
2
2
0
8
7
20
39
from
1B
6
0
0
0
4
2
12
0
5
0
1
4
39
49
2
5
5
29
8
47
96
1 2
11 3
2 0
0 6
7 1
54 26
75 38
2
0
0
0
7
68
77
0
0
0
0
0
10
10
0
1
0
1
9
60
71
0
3
0
11
2
44
58
6
0
0
19
136
94
255
16
37
7
75
185
485
805
direct examination
1C
2
0
0
0
10
0
12
1D
4
0
6
2
12
2
26
1E 2A
3 0
1 0
0 0
0 1
17 9
0 0
21 10
2B
4
0
0
0
15
0
19
2C
0
1
0
0
6
0
7
2D
2
0
0
0
3
0
5
3A
9
1
0
0
1
0
11
3B
9
0
1
0
5
1
16
TOTAL
41
3
7
5
89
6
151
visited
1B
9
2
0
1
0
1
13
1C
2
5
0
0
0
0
7
1D
9
0
0
0
0
0
9
1E 2A
0 0
0 0
0 3
2 0
0 2
0 4
2 9
2B
0
1
2
1
0
8
12
2C
1
0
0
0
0
7
8
20
0
1
5
1
0
21
28
3A
8
0
0
0
0
0
8
3B
13
0
0
0
0
0
13
TOTAL
43
12
10
5
2
41
113
files, and were used to prepare lists for use by field
crews.
The latitude and longitude of each lake were deter-
mined with eleven-point dividers to the nearest de-
gree, minute, and second. The original coordinates
were plotted by computer on transparent map
overlays that were compared to the original maps
to confirm further the coordinates of each lake (Sec-
tion 3.5.1). ID codes were recorded on the topo-
graphic maps for use by field crews in locating the
lakes.
2.3 Applications of the Design
2.3.1 Defining the Target Population
The identity of the map population, subsequently
refined by the exclusion of non-target lakes (Sec-
tion 2.2.4), defines the target population. The num-
ber of lakes comprising the target population is not
known, but rather estimated, because non-target
lakes were identified only in the probability sample.
In contrast, the map population, because of its op-
erational definition as all lakes appearing on
1:250,000-scale USGS maps (Section 2.2.2), is
known precisely and serves as the point of refer-
ence. The map population is larger than the target
10
population by virtue of the exclusion criteria set
forth for non-target lakes. The sample frame is the
structure imposed on the map population by strati-
fication and generation of the list. The frame popu-
lation and sample selection are stratum-specific, as
is the estimation process. Sample data must be
weighted by stratum-specific weights when extrap-
olating from sample results to estimates of popula-
tion parameters. The method of extrapolating to
the population is given in Section 2.3.2 and its im-
portance is discussed in Section 4.1.2.1.
In the broadest sense, the target population is all
those lakes appearing on 1:250,000-scale USGS
topographic maps that could be sampled to yield
data to satisfy the objectives of the ELS-I. However,
the inherent flexibility of the design permits any
subpopulation of lakes from this overall target pop-
ulation to be defined on the basis of any criteria
established from the attributes or measured vari-
ables for any lake. The definition of these still
smaller populations can be determined to answer a
particular question. For example, a subpopulation
of interest can be defined as that set of lakes charac-
terized by elevation greater than 300 m. Or the sub-
population might be defined on the basis of sulfate
or dissolved organic carbon concentrations greater
-------
than or equal to X. Estimates of subpopulation
characteristics can be made in the same manner as
for the entire target population by applying the
proper weights.
2.3.2 Estimating the Target Population Size and
Associated Variance from Sample Data
The process of estimating the target population size
is illustrated in Figure 2-5. The number of target
lakes (N) within an alkalinity map class stratum is
estimated by the equation:
N = N*-ISL
where:
N* = the number of lakes in the stratum map
population, and
Nn = the estimated number of non-target lakes in
the stratum map population.
N* is known from the initial map population (Sec-
tion 2.2.2).
Nn must be calculated from the equation:
Nn = (N*/n*)[nnb + (
Figure 2-5. Estimating the target population size. Eastern Lake Survey -Phase I.
Frame Population
(N*)
M Larae Scale
Taraet ^ MaP
Lakes "" Examination
(Hnb)
hsti mated
Non-Target Lakes
Not Visned Vls'ted
Lakes (n°'
Non- Direct
* larget ^ Examinat.on
1 Lakes
Estimated (n"a)
Non-Target
Lakes
Selected
1 Multiply by N*
n*
Estimated Non-Target
Population Size (Nn)
1
Population Size "~
1 '
Selected Lakes
(n*)
I
Lakes Scheduled for
Visitation
*
Lakes Visited
1
Target Lakes
Visited (n***)
1
Weighting Factor (Wj
W = N/n***
= N*/nf
n' - Effective
Sample Size
(N-N'-Nn)
-------
where:
n* = the total number of lakes in the probability
sample selected from the map population
within the alkalinity map class stratum
nnb = the number of lakes selected in the proba-
bility sample determined to be non-target
by the examination of large-scale maps
nna = the number of lakes in the probability sam-
ple determined to be non-target during
field sampling
q =
n - nnb - n
n - n
nb
= proportion of comple-
tion, i.e., the proportion
of lakes selected for visi-
tation that were actually
visited.
where:
n0 = the number of lakes not visited, represent-
ing incompleteness in the field sampling.
Some of the lakes scheduled for field visitation sub-
sequently were not visited by field crews (Sec-
tion 2.2.4). The group of lakes that were not visited
contained an unknown number of lakes that would
be classified as non-target. It was subsequently
necessary to expand the number of non-target
lakes observed by the field crews (nna) by q to yield
an estimate of the total number of non-target lakes
in the field sample (including the number of non-
target lakes expected within n0).
Weighting or expansion factors are necessary to
estimate the population of lakes in the target popu-
lation and are calculated as follows:
W = N/n*** = N7n'
where:
n*** = n* - nnb - nna - n0, the number of lakes
visited and characterized as target, i.e., the
number of lakes from which water sam-
ples were obtained, and
n' = n*q, the effective sample size, used in
place of n* to account for incompleteness.
Rearranging the equation yields
N = Wn***
The total area of target lakes (A) is estimated simi-
larly:
A = WSAj,
with summation over all lakes in the stratum sam-
ple.
Variances of N and A, for single strata, are esti-
mated by:
VXN) = [N*(N* - n')/(n' - 1)][n***/n'][(n' - rT*)/n'], and
v"(A) = [N*(N* - n')/(n' - 1)][1/n'][SA2 - (SA)2/n'] .
For multiple strata, estimates and variances are ad-
ditive,
ISI = 2Nh, V(N) = SV(Nh), A = 2Ah, and V(A) = 2V(Ah).
Standard errors (of estimates) are the square roots
of variances. (To use the above formulae, see Table
4-4, Section 4.2.2, and calculate n' from n' = N*/W.)
In all analyses, the sample was treated as a simple
random sample within strata. The ordering of the
lakes and the systematic selection process increase
precision over a random sample. Therefore, esti-
mates of precision are conservative.
2.3.3 Subpopulations
Estimations of subpopulation parameters require
more structure than the basic population estimates
and are generally of more scientific interest. For
example, each observed value of X defines a sub-
population of lakes having a value of X s that value
(lakes < 300 m, lakes < pH 6.5, etc.).
Estimates and corresponding variances can be con-
structed for these Subpopulations, over any combi-
nation of strata if the identification of the subpopu-
lation is explicit.
For economy of computation, the algorithms used
in generating all the statistics for the Survey are
quite different, but are mathematically identical and
best understood by formulae generated as a simple
modification of the single stratum equations, given
in Section 2.3.2:
replace each n*
where:
by nz and each 2 by 22
nz = the number of sampled lakes in the
subpopulation z and
2Z = the summation over the sample lakes
in the subpopulation z.
A useful generalization, appropriate for any sub-
population and any combination of strata, is that
N = SW, and A = SWA,
where summation is over the appropriate subset of
sample lakes in the appropriate strata, and where
the weights (W) are assigned according to the stra-
tum in which the lake belongs in the map popula-
tion.
A further generalization, used in the data analyses,
leads to similar formulae for mean and variance of
any variable, X, over any subpopulation and combi-
12
-------
nation of strata:
Mean (X) = 2WX/2W, and Variance
(X) = 2WX2/2W - (2WX/SW)2 .
These are estimates of the parameters of the popu-
lation of lakes defined by the set of sample lakes in
the summation, and characterized by the distribu-
tion, F(X) (Section 4.3.1).
The importance of the weighting factors cannot be
overemphasized. As is shown by example in Sec-
tion 4.1.2, and stated several times in this report (for
example, Section 2.3.2), estimating population
parameters, or examining relationships among
variables with the expectation that these relation-
ships are representative of the population, from
sample data without accounting for weights can
lead to erroneous calculations and incorrect inter-
pretation.
2.3.4 Restrictions
The use and interpretation of any data set are re-
stricted by the design, the quality of the data ob-
tained and the sampling protocols. The frame and
target populations and the period of sampling are
the primary considerations influencing the proper
interpretation of the ELS-I data.
The period of sampling restricts conclusions to the
fall of 1984. Until Phases II and III are completed, the
accuracy of extrapolating the index sample to other
times of the year or to other years will not be
known.
Additional restrictions result from the sampling
frame. Most simply, no conclusions can be drawn
for lake populations not included in the original
frame. Lakes less than 4 ha in surface area are gen-
erally not shown on 1:250,000-scale USGS maps
used to establish the frame population. Conse-
quently, the ELS-I omitted lakes less than 4 ha be-
cause lakes of this size are generally not repre-
sented at this map scale. However, small lakes may
represent an important resource that could not be
sampled practically or statistically in this study.
Therefore, no direct conclusions can be drawn
about this population of lakes.
As was noted in Section 2.3.3, it is essential to the
accurate interpretation of the data to identify clearly
any population for which inferences are made.
However, there is considerable flexibility in defin-
ing a population to be described if it is a subset of
the original target population.
2.4 Lake Characterization
Because of their effects on surface water chemistry,
lake area and elevation, watershed area, and hydro-
logic lake type are important variables to consider
in interpreting chemical data from lakes. These
variables were included in the data base and were
measured or determined as described in the follow-
ing sections.
2.4.1 Lake Area and Elevation
Lake area (in hectares) was measured in duplicate
with an electronic planimeter on 1:24,000-,
1:25,000-, or 1:62,500-scale maps. If lake
boundaries were included on more than one of
these maps, 1:250,000-scale maps were used. Is-
land and swamp areas within the lake shown on the
maps were included in the measurement of water-
shed area (Section 2.4.2) rather than lake area.
When a lake was found to be contiguous with other
bodies of water, divided between adjoining maps,
or exhibited dissimilar shape at different scales of
resolution, the following criteria were used in delin-
eating boundaries:
(1) The boundary of a lake was defined as begin-
ning at the narrowest constriction of its inlet
or outlet.
(2) When a portion of a lake was not shown on an
adjoining map, the perimeter was estimated
using topographic contour lines.
(3) When lakes appeared as two distinct bodies
of water on a 1:250,000-scale map, but as one
on larger-scale maps, the entire water body
was measured as if it were one lake.
The elevation of each lake was recorded in feet
from the largest-scale map available and was sub-
sequently converted to meters. When no lake eleva-
tion was marked on the map, it was interpolated
from topographic contour lines.
2.4.2 Watershed Area
An electronic planimeter was used to measure
watershed area. Watershed areas greaterthan 10 to
15 mi2 (approximately 26 to 39 km2) were measured
on 1:250,000-scale maps. For watersheds less than
10 to 15 mi2, measurements were made from
larger-scale maps.
In some cases, watersheds were not defined clearly
by topography. For example, many lakes in Florida
are located in flat terrain. In such instances, water-
shed boundaries were impossible to determine and
were reported as "undefined" in the data base.
2.4.3 Lake Type
Lakes were classified by hydrologic type (Wetzel
1983) through visual examination of their morphol-
ogy on the largest-scale topographic maps avail-
able. "Seepage" lakes were defined as those lakes
having no inlet or outlet. "Closed" lakes were those
13
-------
with inlets and no outlets. Lakes with outlets but no
inlets or with both were termed "drainage" lakes. A
fourth category comprised artificial lakes or
"reservoirs".
2.5 Water Sample Parameters and
Analytical Methodology
Twenty-five primary and secondary water sample
parameters were selected for measurement on-site
or in analytical laboratories during the ELS-I.
Twenty-four are presented in Table 2-3. Data for
one parameter, base neutralizing capacity, are
being reevaluated (Best et al. 1986) and are not
included here.
Measurements were made in situ, within 16 hours
of sample collection at a field laboratory, or within
7-28 days at an analytical laboratory. Parameters
measured more than once included pH, dissolved
inorganic carbon (DIG) and conductance.
Dissolved inorganic carbon and pH were deter-
mined in the field laboratory using samples sealed
from the atmosphere ("closed system"). Dissolved
inorganic carbon and pH were also measured in the
analytical laboratory on a sample that was exposed
to the atmosphere ("open system"). Additional DIG
and pH measurements were made in the analytical
laboratory on a sample that had been equilibrated
with approximately 300 ppm CO2 in air ("air-
equilibrated").
2.6 Field Methods
The period of field sampling was from October 7,
Table 2-3. Principal Chemical and Physical Measurements, Eastern Lake Survey -Phase I
Parameter3
IN SITU
pH
Conductance
Lake Temperature
Laboratory
Reporting
Units
(iS cm"1
°C
Required
Detection
Limits
—
Intralaboratory
Precision
Goal
—
Maximum Allowable
Sample Holding
Time - Days
(Analytical Lab)
~
Instrument
or Method
Potentiometer Hydrolab
Conductivity cell
Hydrolab
Thermistor
Hydrolab
Reference
(Laboratory
Methods)0
Morris et al. (1986)
Morris et al. (1986)
Morris et al. (1986)
Secchi Disk Transparency m
FIELD LABORATORY
Laboratory pH, closed
system
Dissolved Inorganic mg I
Carbon, closed system
True Color PCU
Turbidity NTU
ANALYTICAL LABORATORY
pH, air-equilibrated
pH, open system —
0.10
Acid Neutralizing
Capacity (ANC)
Aluminum (Al)
extractable
(i-eq
mg L-10.010)
20 (sO.010)
10 (>0.010)
20 (<0.010)
7
7
14
28
Secchi disk
Electrometer Orion
model 611
Infrared
Spectrophotometry
Dohrmann DC-80
carbon analyzer
Comparator Hach
model CO-1
Nephelometer Monitek
model 21
Potentiometer
Potentiometer
Morris et al. (1986)
EPA 150.1
EPA 41 5.2
(modified)
EPA 1 1 0.2
(modified)
EPA 180.1
EPA 150.1
EPA 150.1
Acidimetric Titration Hillman et al. (1986);
modified Gran analysis Kramer (1984)
Atomic Absorption Hillman, et al. (1986)
Spectroscopy, Furnace
on MIBK extract
Atomic Absorption EPA 202.2
Spectroscopy, Furnace
14
-------
Table 2-3. Principal Chemical and Physical Measurements, Eastern Lake Survey -Phase I (Continued)
Intralaboratory Maximum Allowable
Parameter9
Calcium (Ca)
Chloride (Ch)
Dissolved Inorganic
Carbon, air-equilibrated
mgL~1
mgL"1
mgL~1
mg L-"8>
mg I-1"'
mg L-'M
mg L-NS'
mg L-"d>
mg L-'W
mg L-1'8*
mg L-1<8>
mg L-H"'
mg L-W
mg L-"8>
|iS cm-1
0.01
0.01
0.05
0.05
0.1
0.005
0.01
0.01
0.01
0.01
0.01
0.01
0.005
0.002
0.05
0.05
__(f)
Precision Sample Holding
Goal Time - Days Instrument
% RSDb (Analytical Lab) or Method
5
5
10
10
5 (>5.0)
10 (==5.0)
5
10
5
5
10
5
5
10
10 O0.010)
20 (<0.010)
5
5
2
28
28
14
14
14
28
28
28
28
28
28
28
7
28
28
28
14
Atomic Absorption
Spectroscopy, Flame
Ion Chromatography
Infrared
Spectrophotometry
Infrared
Spectrophotometry
infrared
Spectrophotometry
Ion Selective Electrode
Atomic Absorption
Spectroscopy, Flame
Atomic Absorption
Spectroscopy, Flame
Atomic Absorption
Spectroscopy, Flame
Atomic Absorption
Spectroscopy, Flame
Atomic Absorption
Spectroscopy, Flame
Colorimetry (Phenate,
automated)
Ion Chromatography
Colorimetry
(Phosphomolybdate, or
modification,
automated)
Colorimetry
(automated)
Ion Chromatography
Conductivity Cell
Reference
(Laboratory
Methods)0
EPA 215.1
ASTM (1984);
O'Dell et al. (1984)
EPA 41 5.2
(modified)
EPA 41 5.2
(modified)
EPA 41 5.2
EPA 340.2
(modified)
EPA 236.1
EPA 258.1
EPA 242.1
EPA 243.1
EPA 273.1
EPA 350.1
ASTM (1984);
O'Dell et al. (1984)
USGS I-4600-78
USGS I-2700-78
ASTM (1984);
O'Dell et al. (1984)
EPA 120.1
Dissolved ions and metals were determined except where noted.
"Relative precision (% relative standard deviation) calculated for samples at levels above 10 times instrument detection limits, except for pH.
cln situ measurements are outlined in Hillman et al. (1986) and Morris et al. (1986). EPA methods are from U.S. EPA (1983), USGS methods are
from Skougstad et ai. (1979).
dValues converted to (ig L~1 for data analysis. Required Detection Limits are in mg L~1.
"Values converted to M.eq L~1 for data analysis. Required Detection Limits are in mg L"1.
'Mean of six nonconsecutive blank values was required to be less than 9 |xS cm-1.
15
-------
1984 to December 14, 1984. Lakes were accessed
using helicopters with fixed floats. Helicopters
landed on the lake and sampling was conducted
from the pontoons. Samples were transported to a
field laboratory for preliminary processing and
preservation before delivery to the analytical labo-
ratories for analysis within required holding times
(Table 2-3). The activities conducted by each sam-
pling crew are shown in Figure 2-6. Detailed proto-
cols are described in Morris et al. (1986).
Eight sites were selected as field stations, based on
geographic location and the capability to accom-
modate both helicopters and a field laboratory. Re-
mote base sites were established where necessary
to facilitate efficient sample collection. When field
crews operated from remote base sites, samples
were transferred via aircraft to the field station lab-
oratory for processing.
Field stations were staffed by approximately 15
people, including personnel responsible for logis-
tics, helicopter pilots, field sampling crews, and a
field laboratory crew. A two-member sampling
Figure 2-6. Field sampling activities. Eastern Lake Survey -Phase I.
Arrive at
Lake Site
1
Photograph Lake
and Complete Site
Description
1
Collect
Water Samples
I
Make In Situ
Measurements
Site Depth
Secchi Disk Transparency
PH
Conductance
Temperature
Field Blank
Sample
(Deionized
Water)
4-L
Container
Determine
Stratification
Status
Routine
Sample
Two 60-mL
Syringes
4-L
Container
I
Field
Duplicate
Sample
Two 60-mL
Syringes
4-L
Container
Store at 4°C
Complete Field
Data Form
Travel to Next
Lake or Return
to Field
Laboratory
16
-------
team visited five to 10 lakes per day by helicopter.
Generally, two helicopters were assigned to each
field station.
2.6.1 Site Description
Lakes to be sampled were identified from the air
using a LORAN-C guidance system and USGS topo-
graphic maps. Aerial photographs of lakes were
taken. Shoreline disturbances and land use (e.g.,
roads or dwellings) were recorded on a standard-
ized data form. If a lake was accessible by heli-
copter, the aircraft landed near what appeared to be
the deepest portion of the lake. An electronic depth
finder was used to locate more precisely the deep-
est area near where the helicopter landed.
2.6.2 In Situ Measurements
Field measurements, recorded on standardized
field data forms, included site depth, transparency,
pH, temperature, and conductance. Site depth was
determined with a weighted sounding line. Trans-
parency was determined with a black and white
20-cm-diameter Secchi disk. The disk was lowered
on the shaded side of the aircraft within a calm area
between the pontoons.The depths where the disk
disappeared upon lowering and reappeared upon
raising were recorded and averaged.
In situ measurements of pH, temperature, and con-
ductance were determined using Hydrolab model
4041 units that were retrofitted with glass pH elec-
trodes and Beckman Lazarin reference electrodes.
The units were calibrated daily before use with Na-
tional Bureau of Standards (NBS)-traceable pH
buffer solutions (pH = 7.00 and 4.00) and a solution
of 0.001 N KCI (specific conductance = 147 (xS
cm~1). Proper operation of the pH and conductance
functions was checked each day before and after
sampling activities using a CO2-saturated deionized
water solution (theoretical values at STP:
pH = 3.91, specific conductance = approximately
50 |o,S cm"1). The temperature function was
checked daily against an NBS-traceable thermome-
ter.
Measurements using the Hydrolab were usually
made at 1.5 m (Section 2.6.3). A second set of mea-
surements was made at 1.5 m above the lake bot-
tom. If the temperature difference between top and
bottom was <4°C, the lake was classified as non-
stratified. If the temperature difference between
these two depths was s:4°C, a third set of measure-
ments was taken at 60 percent of the site depth. If
the temperature difference between 1.5 m and 60
percent of the site depth was <4°C, the lake was
classified as weakly stratified. If the difference was
>4°C, the lake was classified as strongly stratified.
Temperature and conductance profiles were
recorded for all strongly stratified lakes at either 2-
or 5-m intervals, depending on site depth (Morris et
al. 1986).
2.6.3 Collection of Water Samples
Water samples were collected from a depth of 1.5 m
if the site depth was >3 m using a 6.2-L Van Dorn
acrylic plastic sample bottle (Wildco model 1160-
TT). This sampling depth was arbitrary and was
selected to be below the influence of the helicopter
pontoons and rotor wash. Where the site depth was
<3 m, samples also were collected at 1.5 m if a
clean sample could be obtained. A clean sample
was defined as one free from sediment, plants, or
other large particulate matter. If a clean sample
could not be obtained at 1.5 m, a sample was col-
lected at 0.5 m.
The sampler was modified by installing a nylon
Luer-lok fitting to permit collection of sample
aliquots in syringes without contacting the atmos-
phere. Two 60-mL polyethylene syringes were at-
tached, in turn, to the fitting, filled with water, and
sealed with locking syringe valves. These were the
closed system samples. A bulk water sample was
collected by completely filling a 4-L polyethylene
Cubitainer from the Van Dorn bottle. Both syringe
and bulk water samples were placed in coolers with
frozen chemical refrigerant packs and stored at ap-
proximately 4°C in the dark until processed at the
field laboratory.
Two types of quality assurance samples were col-
lected. Each sampling crew prepared a field blank
sample at the first lake visited each day, using
reagent grade water obtained from the deionized
water system in the field laboratory (Section 2.6.4).
These blanks were prepared by filling the Van Dorn
bottle with deionized water and subsequently
transferring four liters to a Cubitainer. Once daily, a
field duplicate (two syringes and one 4-L Cu-
bitainer) was obtained by collecting a second sam-
ple of water from the same lake.
2.6.4 Field Laboratory Activities
Field laboratory trailers provided a clean and con-
trolled environment in which to process and pre-
serve water samples and to perform certain chemi-
cal analyses as soon as possible after collection.
Each laboratory was equipped with a laminar flow
hood, a deionized water purification and polishing
system, analytical instruments, and equipment for
processing and preserving samples. The laborato-
ries and operations (Figure 2-7) are described in
detail in Morris et al. (1986) and Hillman et al.
(1986).
Upon arrival at each field station, samples were
organized into a field batch for processing. A field
batch was defined as all the samples (including
17
-------
Figure 2-7. Field laboratory activities. Eastern Lake Survey -Phase I.
Field
form
Data Water Audit
s Samples Samples
Samples
^ Organized ^
"~ *• Into Baton -*
for
Processing
i .
Syringes 4-L Containers
4 1
DIC pH Turbidity ^rlje
1 1 1
Field Laboratory Aliquot
Aluminum
Extraction
- (Table ?-3)
1, |,._ _
1 1
* *
Shipped to
Data Management Aliquots 1 -7 Split Samples
Center
Data Base 1
V f
Shipped to Shipped to
Analytical ERL-Corva is,
Laboratory C,anada'
Norway
18
-------
quality assurance samples) that were processed at
a given field laboratory on a given day. Each batch
and sample was given unique identification num-
bers to aid in sample tracking. Batch and sample
numbers and all analytical data obtained at the field
laboratory were recorded on a standardized field
laboratory data form.
Dissolved inorganic carbon was measured using a
Dohrmann DC-80 carbon analyzer on a filtered (0.45
p,m) aliquot from one sealed syringe. The instru-
ment was calibrated daily with a 10.0 mg C L~1
standard and checked during the analysis of each
batch of samples with a 2.0 mg C L~1 solution.
In the field laboratory, pH was measured using an
Orion model 611 pH meter with an Orion Ross
model 81-52 combination pH electrode. The pH
meter allowed for automatic temperature compen-
sation. The pH meter was standardized daily with
NBS-traceable buffer solutions (pH 7.00 and 4.00).
The operation of the meter was checked during the
analysis of each batch of samples with a 10~4 N
H2S04 solution (theoretical pH = 4.00).
The pH of the syringe sample was measured at
ambient temperature with a pH electrode in a flow-
through chamber (Hillman et al. 1986). Because the
syringe was connected to the chamber throughout
the measurements, injection of the sample from the
syringe into this chamber permitted pH to be meas-
ured without exposing the sample to the atmos-
phere. The chamber was rinsed, filled with sample
from the syringe, and the pH reading allowed to
stabilize. Stabilization was defined as a unidirectional
change of 0.02 units or less over a 1 -minute period. A
5-mL aliquot of sample was injected into the chamber.
The pH of this aliquot was recorded when a stable pH
reading was obtained. Additional 5-mL aliquots were
injected and the process repeated until two succes-
sive, stable measurements agreed within 0.03 pH
units. The final pH values was recorded.
Turbidity was determined on subsamples from
each Cubitainer using a Monitek model 21 neph-
elometer. The instrument was calibrated daily with
a 10-nephelometric turbidity unit (NTU) standard.
The nephelometer was checked during the analysis
of each batch of samples with a 5-NTU standard. A
second subsample from each Cubitainer was cen-
trifuged for 10 minutes and analyzed for true color
using a Hach CO-1 color comparator.
Seven aliquots were routinely prepared from each
Cubitainer sample. Table 2-4 summarizes the
preparation and preservation requirements and the
parameters measured in each aliquot by the analyt-
ical laboratories. Certain aliquots were prepared
from filtered (0.45 urn Gelman GA Metricel mem-
brane) portions of each bulk sample. Filtration was
conducted in the laminar flow hood to minimize
potential sample contamination. An aliquot for ex-
tractable aluminum (Al) analysis was prepared from a
portion of filtered sample, using the method described
in Hillman et al. (1986). All extractions were per-
formed within 16 hours of sample collection. Other
aliquots were prepared by pouring samples directly
from each Cubitainer into appropriate containers.
Certain aliquots required preservation with concen-
trated nitric or sulfuric acid (Baker U It rex grade or
equivalent). All aliquots were stored at about 4°C in
the dark until shipment.
Additional aliquots (split samples) were prepared
and preserved from each bulk sample for use in
Table 2-4. Aliquot Preparation and Preservation Requirements, Eastern Lake Survey -Phase I
Aliquot Container8 Preparation Preservation
Parameters
1
2
3
4
5
250-mL Nalgene bottle, acid-leached
10-mL polycarbonate centrifuge
tube
250-mL Nalgene bottle, leached
with deionized water
125-mL Nalgene bottle, acid-leached
500-mL Nalgene bottle, leached
Filtered, acid-rinsed filtration unit
Filtered, 8-hydroxy-quinoline
complexatipn, MIBK extraction
Filtered, filtration unit not acid-
rinsed
Filtered, acid-rinsed filtration unit
Unfiltered
Acidified with HN03
to pH <2; 4°C
4°C
4°C
Acidified with H2SO4
to pH <2; 4°C
4°C
Ca, Mg, Na,
K, Mn, Fa
Extractable Al
cr, so,-2,
N03', F-, Si02
DOC, NH4*
DIG, pH, con-
with deionized water
125-mL Nalgene bottle, acid-leached Unfiltered
125-mL Nalgene bottle, acid-leached Unfiltered
Acidified with H2SO4
to pH <2; 4°C
Acidified with HN03
to pH <2; 4°C
ductance,
ANC, BNC
Total P
Total Al
aAII Nalgene made of high-density linear polyethylene.
19
-------
data comparability studies. Splits from all samples
were sent to the EPA Environmental Research Lab-
oratory in Corvallis, Oregon (ERL-Corvallis), for ele-
mental analysis. Other splits were sent to analytical
laboratories in Norway and Canada. The results of
these analyses will be described in a future report.
Following preparation and preservation, all sample
aliquots were refrigerated overnight at 4°C. The fol-
lowing morning, samples were packed in insulated
shipping containers with sufficient packages of
frozen chemical refrigerants to maintain a tempera-
ture of about 4°C during shipment. Samples were
shipped to the appropriate analytical laboratory by
overnight courier or commercial air freight. A ship-
ping form was used to track samples from the field
laboratory to the analytical laboratory. Aliquots ar-
rived at the analytical laboratories within 54 hours
of collection. Copies of field and field laboratory
data forms were shipped to Oak Ridge National
Laboratory (ORNL) in Oak Ridge, Tennessee, for
data entry, and to quality assurance staff at the EPA
Environmental Monitoring Systems Laboratory in
Las Vegas, Nevada (EMSL-Las Vegas). Copies also
were sent to the EPA Sample Management Office
for sample and data tracking. One copy of each
form was retained at the field laboratory.
2.7 Analytical Support
Because of the large number of samples and the
required holding times, the use of a single labora-
tory for sample analysis was not possible. For this
reason, detailed criteria were prepared that defined
the analytical and quality assurance requirements
and bids were solicited from analytical laboratories.
Prospective laboratories were evaluated through
the analysis of performance evaluation samples
(Drouse et al. 1986) and on-site inspections.
All analyses were conducted according to handling,
analytical, and quality assurance protocols detailed
in Hillman et al. (1986) and Drouse et al. (1986). Four
analytical laboratories (Table 2-5) were selected.
The goal was to analyze all samples collected
within one region at the same laboratory; however,
due to logistical problems, two laboratories were
needed to analyze samples from each region.
2.8 Data Base Management
2.8.1 Overview
The objectives of the data management component
of the ELS-I were to: (1) enter the data, (2) prepare
a validated data set, and (3) perform statistical anal-
ysis and evaluation of the data. The Environmental
Sciences Division of ORNL designed and imple-
mented data base management activities, using the
Statistical Analysis System software (SAS Institute
1982) on tandem IBM 3033 mainframe computers.
Table 2-5.
Laboratory
Distribution of Lake Samples to Analytical Labo-
ratories, Eastern Lake Survey -Phase I.
Number
of Lakes
Sampled8
(% of
Total)
Subregion
(% of
Total)
Global Geochemistry Corp.
Canoga Park, CA
Rockwell International
(Environmental Monitoring
Services, Inc.),
Thousand Oaks, CA
Versar, Inc.
Springfield, VA
238 (13.2)
752(41.8)
649(36.1)
U.S. Geological Survey
Arvada, CO
159 (8.8)
2A (12.0)
2B (28.8)
2C (20.3)
3A (100)
1A (99.5)
1B (50.0)
1C (99.5)
1D (59.7)
1E (100)
1A (0.5)
1B (50.0)
1C (0.5)
1D (40.3)
2A (88.0)
2B (71.2)
2C (79.7)
3B (100)
alncludes regular and special interest lakes.
The validated data set consists of individual records
for each lake sampled and contains site descriptors,
field observations, laboratory measurements, and
calculated variables for the lakes sampled. The defi-
nition of all variables measured with their formats,
units of measure, and comments is presented in a
data base dictionary (Kanciruk et al. 1986b). De-
tailed discussions of the quality assurance meas-
ures developed for the data base are given in Rosen
and Kanciruk (1985) and Kanciruk et al. (1986a).
2.8.2 Data Base Design and Data Flow
The lake ID assigned before field work began (Sec-
tion 2.2.6) was used as the key site identifier. The ID
numbers were cross-referenced to lake names,
state and county Federal Information Processing
Standards codes, latitude and longitude, and other
identifiers within the site descriptor file.
Data base development is summarized in Figure
2-8. The working data base consisted of three main
data sets: raw (Data Set 1), verified (Data Set 2), and
validated (Data Set 3). Each contained numerous
SAS relational (tabular) data files (SAS Institute
1982). The verified and validated data sets were
developed from the raw data set as quality assur-
ance procedures were implemented (Section 3).
The final reported data set (Data Set 4) consists of
the validated data set, modified after removing er-
roneous data and substituting for missing values
when appropriate (Section 3.4). Protocols used for
such changes are discussed in Eilers et al. (1986).
20
-------
Figure 2-8. Data base development. Eastern Lake Survey--Phase I
verified Data Set
(Data Set 2)
Substitution
and
Replacement
i
Final ELS-1 Data Set
(Data Set 4)
Validated Data Set
(Data Set 3}
21
-------
Copies of all data sets have been maintained on
tape as a permanent record.
Background information (lake names, ID numbers
and physical characteristics) was transmitted from
ERL-Corvallis to ORNL, where it was entered into a
data set. Data forms with the information obtained
at field sites and analytical laboratories were sent to
ORNL and to EMSL-Las Vegas. The information
was entered into the data base at ORNL After entry
the data were sent to the EPA IBM computer at
Research Triangle Park, North Carolina, for access
and review at EMSL-Las Vegas. Corrections and
flags were returned to ORNL and entered into the
verified data sets. After verification, the data were
validated jointly by ERL-Corvallis, EMSL-Las Vegas
and ORNL staff.
2.8.3 Data Base Structure
The final data set consists of three relational data
files, one each for the regular lakes and special in-
terest lakes and one for all sampled lakes. The infor-
mation pertaining to each lake, including all back-
ground information such as latitude and longitude,
country, elevation, as well as the measured vari-
ables, has been merged into one observation per
lake. In addition, a program for determining the
population weighting factors for each stratum (Sec-
tion 2.3.2) is included.
22
-------
Section 3
Quality Assurance
Quality assurance (QA) and quality control (QC)
procedures for sampling and field and analytical
laboratory operations were implemented for the
Eastern Lake Survey-Phase I (ELS-I). Development
of these procedures required preparation for sam-
pling, implementation of QA/QC activities during
field operations, and verification and validation of
the data after sample analysis.
3.1 Preparation for the Survey
In preparation for lake sampling, a draft QA plan
and draft field and laboratory protocols were writ-
ten and later evaluated during pilot studies (prelim-
inary field studies). The results of these pilot stud-
ies were used to evaluate all aspects of the Phase I
plans and protocols. The draft plans and protocols
were revised based on the results of the pilot stud-
ies. The QA plan was developed prior to implemen-
tation of the Survey (Drouse et al. 1986). The final
field sampling and laboratory protocols were pro-
vided to all field personnel. Intensive training was
also provided for the field personnel prior to sam-
pling.
3.2 Implementation of Quality
Assurance/Quality Control Activities
Several QA/QC activities were implemented during
field operations. All field stations and analytical lab-
oratories were visited by QA personnel during field
operations to evaluate performance and ensure
that protocols were being followed. Quality control
samples (Table 3-1) were used to assure that instru-
ments and data gathering activities were operating
withing limits established by the QA plan. Quality
assurance samples (Section 3.5.2) were used to
evaluate the performance of field and analytical
laboratories and to establish precision estimates. A
number of the QA samples were added to the
batches in the field and submitted to the analytical
laboratories with lake samples. Neither the sample
identification nor its composition was disclosed to
the analytical laboratory. The relationship of these
QA/QC samples to the processing of the lake sam-
ples is illustrated in Figure 3-1. Daily contact with
the analytical laboratories was made by the QA
staff to ensure that appropriate QC protocols were
being implemented, and to identify and correct
problems if and when they occurred in the labora-
tory.
The sample data, including QA/QC information,
were entered into the raw data set. Data were en-
tered into two separate data files independently. A
computer program was developed to compare the
two files and identify input errors. The data entry
procedure had an estimated error rate of 1 in 6000,
or 0.017 percent. After this QC step was completed,
errors were corrected and the resulting data were
merged into the raw data set.
Table 3-1. Descriptions and Applications of Quality Control
Samples, Eastern Lake Survey -Phase I
Sample Type Descriptions Application Frequency
Trailer
Duplicate
Lake sample;
split
Field lab; control
within-batch
precision
One per field
batch
Laboratory
Blank
Matrix Spike
Quality
Control
Check
Analytical
Laboratory
Duplicate
Zero analyte
standard
Batch sample
plus known
quantity of
analyte
Standard;
source other
than calibra-
tion standard
Sample
aliquot; split
Field and analyt-
ical lab; control
signal drift and
sample contami-
nation
Analytical lab;
control sample
matrix effect on
analysis
Field and analyt-
ical labs; control
accuracy and
consistency of
calibration
Analytical lab;
control within-
batch precision
One per
laboratory
batch
One per
laboratory
batch
Before, after
every 10,
and after
final sample
One per
laboratory
batch
3.3 Data Verification and Validation
Verification of the raw data involved an extensive
review of the reported sample data and associated
QA/QC information. This review process was used
to identify questionable data and to correct, qualify,
or eliminate individual values if necesary. The veri-
fied data set was validated during data analysis.
Details of the verification and validation procedures
follow.
23
-------
Figure 3-1. Collection and processing QA and QC samples. Eastern Lake Survey -Phase I.
Field Sampling
Routine Lake
Samples
Field Blank
Samples
1
Field Duplicate
Sample
I
Field/Laboratory
Audit Sample(s)
Field Laboratory
~\
Daily Batch
of Samples
Relabeling
Analysis
(DIG, pH, Turbidity, True Color)
T
QC Check
Samples
f
Batch
Samples
-»• Data -
f
Trailer
Duplicate
Sample
I
Aliquot Preparation
Aluminum Extraction
Preservation
Shipment to
Analytical Laboratory
Analytical Laboratory
— "I
Analysis
r
QC Samples
Batch
Samples
Data
3.3.1 Verification
Data verification was accomplished by establishing
a systematic process to identify and review ques-
tionable values in the raw data set (Figure 3-2). Data
packages, consisting of analytical results and sup-
porting information from the field and analytical
laboratories, were sent simultaneously to Oak
Ridge National Laboratory (ORNL) for data entry
and to the EPA Environmental Monitoring Systems
Laboratory (EMSL-Las Vegas) for preliminary re-
view. At these locations, data were checked for
24
completeness and acceptability. Obvious problems
with the analytical data were referred to the analyt-
ical laboratory for confirmation, correction, or re-
analysis of the samples.
After entry into the raw data set, the data were
verified on a sample by sample basis. To be veri-
fied, data from the analysis of a lake sample had to
meet acceptance criteria (Drouse et al. 1986) for
both anion-cation balance and percent difference
between measured and calculated conductance.
Discrepancies were flagged unless they could be
-------
Figure 3-2. Data verification procedures. Eastern Lake Survey -Phase I.
Batch Data
Report Complete
Data Files Compared,
Corrections or Flags
Applied
Data Mgmt. Center
Applies Corrections
Yes
No
Correct Reporting
Errors, Set Flags
Data Mgmt. Center Applies
Corrections and Flags into
Data Set
25
-------
corrected by adjusting a reporting error or could be
explained by the presence of dissolved organic pro-
tolytes. A protolyte analysis program (Drouse et al.
1986) was used to estimate the relative contribu-
tions of carbonates and organic protolytes to acid
neutralizing capacity (ANC). Data values for a sam-
ple were also flagged when they did not meet ac-
ceptance criteria developed from the analysis of
external QA samples (Section 3.5.2), internal QC
checks, or when prescribed sample holding times
were exceeded.
Suspected analytical errors were referred to the an-
alytical laboratory for reanalysis. Data from reanal-
ysis were evaluated (Drouse et al. 1986) and accept-
able values were flagged and substituted for the
original values in the verified data set.
Less than three percent of the raw data reported for
lake samples was classified as exceptions and re-
quired correction before transfer to the verified
data set. Sample reanalysis was requested for less
than one percent of the originally reported raw data
values. Less than one percent of the reported data
required correction because of transcription or data
entry errors
3.3.2 Validation
The data validation process (Figure 3-3) identified
possible errors in chemical analyses that could not
be revealed by verification procedures (Eilers et al.
1986). The validity of non-chemical measurements
was also evaluated.
Observations that were not typical of other sample
values, i.e., outliers, were detected using a variety
of approaches. Initially, each variable was consid-
ered individually to identify values that were out-
liers with respect to the sample distribution. Box-
plots (Tukey 1977) for each variable were prepared
using statistical procedures (SAS Institute 1982).
These plots summarized the data for a variable
using the difference between the upper and lower
quartile (interquartile range). For this procedure,
outliers were arbitrarily defined as those values
greater than the absolute value of three times the
interquartile range.
Certain pairs of variables were expected to exhibit
a linear relationship. Outliers in these relationships
were identified by examination of scatter plots anc
least squares linear regression analyses. Standard-
Figure 3-3. Data validation procedures. Eastern Lake Survey -Phase I
Yes
Remove All
Flags Except
for Nitrate
Add Validation
Flag; Retain
Verification
Flag
26
-------
ized residual values were calculated using the fol-
lowing formula:
{actual value - predicted value)
[residual standard deviation!
Values greater than three were identified arbitrarily
as outliers. Because least squares analysis can be
strongly biased by certain types of outliers (Velle-
man and Hoaglin 1981), the residuals from resistant
line fits (i.e., lines fit through the medians of parti-
tions of data) were examined for selected variables
measured in the field and field laboratory. Analyti-
cal variables were examined using an iterative
process of linear regression to identify additional
outliers that would not have been identified without
previously removing the major outliers. Outliers
among related groups of variables were detected
using the SAS FASTCLUS cluster analysis proce-
dure (SAS Institute 1982) and principal components
analyses. Possible outliers associated with major
cations and anions were examined using trilinear
diagrams (Hem 1970). Outliers were screened for
agreement with other variables that might explain
their high residual standard deviation. Those out-
liers remaining after confirmation of the reported
data were flagged in the validated data set. In a
limited number of cases, sufficient evidence was
available to indicate that the reported value was in
error. These values were flagged in the validated
set and marked for substitution in the final data set.
Analytical variables having possible systematic er-
rors were detected by comparing values from the
ELS-I to those from other lake survey data sources
(Eilers et al. 1986). These data sets were selected on
the basis of geographic location, accessibility of the
data, and documentation of QA procedures. Com-
parisons were made using scatter plots and linear
regression procedures to identify values requiring
additional scrutiny. Systematic differences were
not sufficiently large to indicate analytical errors of
a magnitude that would affect data analysis and
interpretation.
3.4 Development of Final Data Set
The calculation of population estimates (Section
2.3.2) is difficult if values are missing from the data
set. A final data set (Data Set 4) was prepared to
resolve problems in the validated data set resulting
from missing values. Data Set 4 also was modified
by averaging field duplicate values and substituting
for analytical values determined to be in error dur-
ing validation (Figure 3-4).
Substitution for missing values was done in one of
several possible ways. Values from duplicate sam-
Figure 3-4. Development of Data Set 4, Eastern Lake Survey-Phase I.
/Validated /
/ Data Set /
Flagged
for
Substitution
Unflagged
Duplicate
Available
Unflagged
Duplicate
Available
Use
Average of
Routine and
Duplicate
(Flag)
Good
Relationship
with Other
Variable(s)
Use
Simple/Multiple
Regression
to Find Substitute
(Flag)
27
-------
pies were used when available. Redundant analy-
ses were performed for pH, DIG, and conductance
(Section 2.5). Redundant measurements on split
samples (Section 2.6.4) were performed for metals
and other elements. If a duplicate measurement
was not available, a comparable measurement was
chosen and substituted for the missing value using
a linear regression routine. If redundant measure-
ments were not available or acceptable, a substitu-
tion value was calculated from the available data
using observed relationships with other variables
(e.g., sodium and chloride). The last option for iden-
tifying a substitution value was to use the stratum
mean. All substitution values were examined a sec-
ond time for acceptability before including them in
the final data set. Substituted values were flagged
as such in the final data set. A total of 145 values of
approximately 45,000 chemical measurements ex-
amined during the validation process was substi-
tuted in the final data set. Of these 145 values, 15
were missing from the data collection process and
130 values were substituted for suspected meas-
urement errors.
Two other changes were made in the final data set.
If duplicate data met QA precision criteria, the aver-
age of the duplicates was used in the final data set.
Negative values for parameters other than ANC,
that resulted from analytical calibration bias, were
set equal to zero. The bias in the estimate of vari-
ance due to this adjustment did not affect data
analyses. All values modified in the final data set
were flagged.
3.5 Quality Assurance/Quality Control
Results
3.5.1 Site Confirmation and Characterization
Confirmation that the sampled lakes were the in-
tended lakes was accomplished by comparing
aerial photographs taken during the Survey to top-
ographic maps. Five lakes were mistakenly sam-
pled. The data from these lakes were deleted from
the final data set and the lakes classified as "not
visited" (Section 2.2.4). Five additional samples
were also removed from the data base. One lake,
Quabbin Reservoir, was sampled twice; only the
sample from the major basin is retained in the files.
Two lakes were sampled with conductance values
exceeding the maximum criterion of 1500 u,S cm~1.
Two additional unnamed lakes were identified as
target lakes on the maps, were sampled by the field
crews, but were later determined to be affected by
discharges from waste outfalls.
3.5.2 Evaluation of QA Sample Data
Data from the analysis of QA samples from the
ELS-I (Table 3-2) were used to estimate the quality
28
of the analytical measurements (Best et al. 1986).
Field blanks were used to estimate the background
value for each parameter. The system decision limit
(SDL) represents the lowest measured sample
value that can be distinguished from blanks. For the
purpose of this report, this value was chosen to be
the 95th percentile of the distribution of field blank
measurements (Table 3-3). The SDL should not be
confused with the laboratory detection limit. The
latter evaluates only the analytical method detec-
tion limit while the SDL also includes potential con-
tamination from sampling, processing and ship-
ping.
The units for cations (calcium, magnesium,
sodium, potassium, and ammonium) and anions
(sulfate, nitrate, chloride, and fluoride) are pre-
sented in both concentration units as reported from
the laboratories (mg L~1) and the units used for
data analysis (ixeq L~1), shown in parentheses.
Field blanks were also used to calculate the quanti-
tation limit, which was ten times the standard devi-
ation of field blank measurements. The quantitation
limit defines the concentration of an analyte that is
high enough to be meaningful for estimating the
precision of the sampling and analysis processes.
Table 3-2. Descriptions and Applications of Quality Assur-
ance Samples, Eastern Lake Survey -Phase I.
Sample Type Description
Application
Frequency
Field Blank Deionized Estimate system One per
water (ASTM decision limit sampling
Type 1) and quantitation crew per day
treated as a limit
lake sample
Laboratory Zero analyte Identify sample One per
Blank standard contamination laboratory
batch
Field
Duplicate
Trailer Dupli-
cate
Laboratory
Duplicate
Field Audit
Laboratory
Audit
Duplicate
lake sample
Lake sample;
split
Sample
aliquot; split
Synthetic
samples,
natural lake
samples
Synthetic
samples
Estimate overall
within-batch pre-
cision
Estimate analyti-
cal within-batch
precision
Estimate analyti-
cal within-batch
precision
Estimate overall
among-batch
precision; esti-
mate laboratory
bias
Estimate analyti-
cal among-batch
precision; esti-
mate laboratory
bias
One per field
station per
day
One per field
batch
One per
laboratory
batch
A minimum
of one field
or laboratory
audit per
field batch
A minimum
of one field
or laboratory
audit per
field batch
-------
Table 3-3. Evaluation of Field Blank Data, Eastern Lake Sur-
vey-Phase I
Parameter (Unit)
ANC (jxeq L~1)
Conductance (|xS cm~1)
DIC, open system (mg L"1)
DIC, air-equilibrated (mg L"1)
DOC (mg L'1)
Ca (mg L~1)
Fe (p,g L-1)
K(mgL"1)
Mg (mg L"1)
Mn (M.Q L-1)
Na (mg L~1)
Cr (mg L-1)
SO4"2 (mg/L"1)
N03" (mg/L"1)
F~, total dissolved (mg L~')
NHU* (mg L"1)
Al, extractable (|xg L~1)
Al, total fug L-')
Si02 (mg L-1)
P, total (pig L~1)
System Decision
Limit8 (n = 245)
6.9
1.3
0.42
0.28
0.4
0.03(1.5)
22
0.02 (0.4)
0.01 (0.7)
11
0.03 (1.2)
0.08 (2.3)
0.09(1.9)
0.389(6.3)"
0.023 (0.4)c
0.919 (14.8)d
0.005 (0.3)
0.04(2.1)
8
30
0.11
8
aValues in parentheses are in |xeq L '.
bValue is for analyses of aliquot 3 after protocol change and
aliquot 5 before protocol change (n = 245, Section 3.5.2).
"Value is for analyses of aliquot 3 after protocol change (n = 99,
Section 3.5.2).
"Value is for analyses of aliquot 5 (n = 146, Section 3.5.2).
Measurement of sample concentrations less than
this value will have higher variability or poorer pre-
cision. Because variability increases substantially
as the concentration approaches zero, only values
above the quantitation limit were used to estimate
precision.
Laboratory blank data also were used to judge ana-
lytical laboratory performance (Drouse et al. 1986).
Laboratory blank values were compared to field
blank measurements to identify possible sample
contamination that occurred during sample collec-
tion, processing, and shipment. The reported in-
strumental detection limits (Best et al. 1986) indi-
cated that the required detection limit (RDL) was
achieved for all parameters except iron. Reanalysis
for iron was requested for those samples where the
RDL was not achieved. This reanalysis was per-
formed by the alternate graphite furnace method
(Hillman et al. 1986), which had an acceptable
detection limit.
Data from the analyses pf 125 field duplicate sam-
ple pairs collected during the ELS-I were used to
estimate overall within-batch precision, which in-
cluded the variability introduced as a result of sam-
ple collection, processing, and analysis. Analytical
within-batch precision was also estimated using
118 trailer duplicate pairs and 127 laboratory dupli-
cate pairs. The estimates of overall and analytical
within-batch precision did not include the effect of
among-batch variation that may have been caused
by day-to-day differences such as different calibra-
tion curves.
Estimates of overall and analytical within-batch
precision for field laboratory and analytical labora-
tory measurements are presented in Table 3-4. Pre-
cision estimates for field laboratory measurements
were calculated using all available duplicate data,
regardless of measurement value. Precision esti-
mates for measurements performed in the analyti-
cal laboratories were calculated using duplicate
pairs having mean values greater than the quantita-
tion limit. Quantitation limits could not be calcu-
lated for field laboratory measurements because
field blanks were not measured in the field labora-
tory; and therefore, some values less than the
quantitation limit were included.
For all four field laboratory measurements (Table
3-4), the analytical within-batch precision was bet-
ter than the intralaboratory precision goals estab-
lished for the ELS-I. Overall within-batch precision
of the pH measurements was also better than ex-
pected intralaboratory goals. Overall precision esti-
mates for DIC and turbidity measurements were
within expectations for interlaboratory precision
(estimated as twice the value of intralaboratory pre-
cision). Analytical within-batch precision estimates
(pooled across all analytical laboratories) were bet-
ter than the intralaboratory precision goal for all
measurements except air-equilibrated pH, conduc-
tance, chloride, sulfate, and extractable aluminum.
The precision for air-equilibrated pH, extractable
aluminum and sulfate was within the expectation
for interlaboratory precision (estimated as twice the
intralaboratory precision goal). With the exception
of chloride, DOC (>5 mg L~1), nitrate, and total alu-
minum (>0.010 mg L~1), overall within-batch preci-
sion for measurements made in the analytical labo-
ratories w'as within the expectation for
interlaboratory precision.
Among-batch (interlaboratory) precision cannot be
estimated from the field duplicate and analytical
laboratory duplicate data because these pairs were
not divided among the analytical laboratories.
However, audit sample data can be used to esti-
mate interlaboratory precision. Table 3-5 presents
the among-batch precision estimated from 41 field
natural audit samples. A detailed evaluation of
among-batch precision for the ELS-I is presented in
Best et al. (1986).
During the early part of the ELS-I, unacceptable lev-
els of nitrate were reported for field blank samples,
resulting in non-systematic bias in the nitrate analy-
ses. This contamination was traced to the field lab-
oratory. It was attributed to residual nitric acid from
29
-------
Table 3-4. Estimated Within-Batch Precision from Field, Trailer and Laboratory Duplicate Data, Eastern Lake Survey -Phase I
Parameter
Intralaboratory
Precision Goal
%RSD"
Overall Within-Batch Precision
(Field Duplicates)
n" RMS of %RSD°
Analytical Within-Batch Precision
(Trailer/Laboratory Duplicates)
nb RMS of %RSDc
Field Laboratory
pH, closed system
DIG, closed system
True Color
Turbidity
Analytical Laboratory
0.1d
10
10
10
124e
123e
125e
125e
0.04
16
22
19
93e
116e
118e
117e
0.01
4.6
1.5
8.4
pH, open system
(initial ANC)
pH, air equilibrated
ANC
Conductance
DIG, open system
DIC, air equilibrated
DOC
(s5 mg L-1)
(>5 mg L-1)
Ca
Fe
K
Mg
Mn
Na
cr
S04~2(mg/L-1)
N03~(mg/L~1)
F~, total dissolved
NH4+(mgL-1)
Al, extractable
O0.10 mg L~1)
(<0.10 mg L-1)
Al, total
O0.10 mg L-1)
(sO.10 mg L"1)
SiO2
P, total
O0.10 mg L-1)
(<0.10 mg L-1)
0.05d
0.05d
10
1
10
10
10
5
5
10
5
5
10
5
5
5
108
10
10
5
5
10
20
10
20
5
10
20
125e
125e
90
125
85
94
46
59
125
32
82
125
6
121
85
115
5
9
0
62
0
9
0
1
0
50
4
0
0.05
0.09
10
1.9
3.7
5.0
5.6
12
2.3
10
3.7
2.3
11
4.3
17
6.5
60
45
—
8.9
h
11
—
24'
—
2.7
9.7
—
127e
127e
86f
123
113
94
73
41
123
101
121
121
73
121
124
124
86
39
46
123
54
45
0
9
0
64
95
9
0.04
0.08
2.1f
10
3.5
2.2
2.5
2.4
0.88
4.3
1.5
0.64
1.7
0.96
7.1
11
3.6
3.7
3.3
2.5
2.3
18
—
5.4
—
2.2
10
8.6
aRSD = relative standard deviation.
bn = number of duplicate pairs with mean > quantitation limit (Best et al. 1986).
Calculated as the root mean square (RMS) of the relative standard deviation of the duplicate pairs (for pH, RMS of the absolute
standard deviation was calculated).
dAbsolute standard deviation (pH unit).
en = total number of duplicate pairs.
'Analytical laboratories were not required to analyze calibration blanks for ANC. Values for routine and duplicate pairs with means
greater than the quantitation limit (56.6 (xeq L"1) were calibrated using field blanks.
9The first value is for analyses of aliquot 5 and aliquot 3 after the protocol change; the second value is for aliquot 3 only; and the third
for aliquot 5 only (Section 3.5.2).
hAII pairs had mean values s quantitation limit. For all pairs with mean >0 (n = 113), the precision was 34%.
'Only pair had mean quantitation limit; therefore, the calculated precision (RMS of %RSD) for that pair was equivalent to the
actual %RSD. For all pairs with mean >0 (n = 125), the precision was 35%.
30
-------
the solution used to rinse the filtration apparatus
prior to sample preparation. To eliminate this
source of contamination, the procedure was
changed to use a separate filtration unit to collect
aliquot 3 (Section 2.6.4, Table 2-4) for anion analy-
ses. This filtration unit was not rinsed with nitric
acid at any time. Nitrate concentrations measured
in field blanks prepared using the modified proce-
dure were <0.2 n,eq L~1. Nitrate data from samples
analyzed before the modified protocol was imple-
mented were not used. They were replaced with
values obtained by analyzing the corresponding
aliquot 5 samples (which were not filtered or pre-
served with acid) for nitrate after filtration of the
that were identified and resolved through the im-
plementation of the QA program are also summa-
rized in that report.
Table 3-5. Among-Batch Precision Estimated from Field
Natural Audit Samples, Eastern Lake Survey-
Phase I
Parameter (Unit)
Mean Concentration8 Standard
(n = 41) Deviation
pH, open system (initial ANC)
pH, air-equilibrated
ANC (ixeq L"1)
Conductance (jiS crrr1)
DIC, open system (mg L~1)
DIC, air equilibrated (mg L~1)
DOC (mg L-')
Ca (mg L-1)
Fe (n.g L-i)
K (mg L-1)
Mg (mg L'1)
Mn (ixg L-')
Na (mg L~1)
Cl- (mg L-1)
S04'2(mgl_-1)
NO3" (mg L"1)
F~, total dissolved (mg L~1)
NH4+(mgL-1)
Al, extractable (jig L"1)
Al, total ((jig L-')
Si02 (mg L~1)
P, total (fig L~1)
5.07
5.18
(2.36)b
26.7
0.42"
0.1 9b
3.3
1.91 (95)
21"
0.49 (13)
0.35 (29)
70b
0.67 (29)
0.61 (17)
6.95 (145)
1.425 (23)c
1.473 (24)d
1.403 (23)e
0.077 (4)
0.06 (3)b
182
305
4.33
2b (0.0016)
0.04
0.27
6.1
1
0.11
0.11
0.3
0.2
14 (13.8)
0.02
0.01
18
0.03
0.31
0.51
0.08C
0.04d
0.078
0.003
0.03
62
85
0.35
2 (0.0023)
aValues in parentheses are in (xeq L"1.
"Values are less than quantitation limit (Best et al. 1986).
cValues are for analyses of aliquot 3 after protocol change and
aliquot 5 before protocol change (n = 41, Section 3.5.2).
"Values are for analyses of aliquot 3 (n = 13) after protocol change
(Section 3.5.2).
"Values are for analyses of aliquot 5 (n = 27, Section 3.5.2).
samples at the analytical laboratory. Table 3-3 pro-
vides decision limits for nitrate as the variable ap-
pears in the data base for both the reanalyzed sam-
ples (from aliquot 5) and samples analyzed after the
protocol was modified (aliquot 3). The second value
given is for analysis of aliquot 3 after corrective
action was taken. The third value shown is for anal-
ysis of aliquot 5. Results of QA sample evaluation
for nitrate are presented in Best et al. (1986). Other
minor problems involving sample contamination
31
-------
Section 4
Results of Population Estimates
4.1 Data Presentation and
Considerations
4.1.1 Presentation
In the interest of condensing the results, all tables
except those describing the sample and target pop-
ulation (Section 4.2.3) show combined results at the
regional and subregional population levels, omit-
ting the results of the alkalinity map class strata
(Section 2.2.1). Summaries of regional population
estimates are provided for Regions 1 and 2, but not
for Region 3. The lakes in Subregions 3A and 3B
were found to be so dissimilar that it would be
misleading to combine the results from these sub-
regions to obtain regional estimates. For the same
reason, population estimates for regions are' not
combined.
Only the primary variables (pH, ANC, sulfate, cal-
cium, extractable Al and DOC) are summarized by
extensive tables in this section. Of the variables
measured, these were selected because of their di-
rect relevance to issues of acidic deposition effects.
Lake population statistics are also provided for esti-
mates of the physical attributes of the lakes sam-
pled, as well as for secondary variables of interest.
Relationships among variables are discussed in
Section 5. Volume II of this report gives summary
statistics on all the variables measured.
The results of the Survey describe the chemical
status of the lakes by comparing cumulative fre-
quency distributions of lake populations (Section
4.3) among and within regions. An assumption im-
plicit in these analyses is that chemical differences
among populations ultimately can be attributed to:
(1) characteristics of the watershed; and (2) the
chemistry of deposition. However, lakes also are
chemically altered by anthropogenic disturbances
that cannot be ascertained from evaluation of
watershed land use types. For example, addition of
lime Ca(HCO3)2 to a lake will dramatically modify
the lake ANC, but the nature of the alteration may
be undetected in a synoptic survey. Modification of
lake chemistry as a result of liming is not random
among subregions, being performed predomi-
nantly on (formerly) acidic lakes. The magnitude of
this and other confounding effects is not known.
4.7.2 Design Considerations
4.1.2.1 Design Constraints
The design of the ELS-I requires that the results be
presented as population and/or subpopulation esti-
mates whenever conclusions combining strata are
to be drawn (Section 2.3.3). As noted throughout
this report, expansion factors or weights (W) must
be used when making combined strata estimates of
attributes for the population of lakes. These
weights are defined, and the estimating aquations
are given, in Section 2.3. To emphasize the signifi-
cance of the cautionary notes relating to weighting,
Table 4-1 was developed as an example to compare
sample results to estimated population results for
pH and ANC. An examination of this table demon-
strates the significance of the weighting factors in
interpreting the findings of the ELS-I. The ANC val-
ues in Subregion 2D particularly exemplify the er-
rors that can result from using unweighted sample
data to infer estimated population medians.
Table 4-1.
Comparison of Sample Median pH and ANC to
Estimated Population Medians Using Weighting
Factors, Eastern Lake Survey- -Phase I
PH
ANC ((Jieq L"1)
Estimated
Sample Population
Median Median
Estimated
Sample Population
Median Median
SUBREGION
1A
1B
1C
1D
1E
6.74
6.77
6.78
6.71
6.89
6.71
7.02
6.77
6.81
6.91
116
145
124
131
138
112
297
120
162
148
REGION
1
SUBREGION 2A
2B
2C
2D
REGION 2
SUBREGION 3A
3B
6.80
6.83
6.61
6.47
6.76
6.74
6.93
6.52
6.87
6.94
7.10
6.68
7.39
7.09
6.98
6.56
129
145
97
55
134
120
202
73
158
185
284
94
802
360
250
84
32
-------
Another example, using strata 2A1 and 2A2, illus-
trates the requirement that all estimates be made
within strata, and that mean or other statistics in-
volving more than one stratum be calculated with
the appropriate stratum weights (Table 4-2). The
correct way to estimate the total number of lakes
below a reference value (in this example pH <6.0)
in two strata is to determine first the total number
of lakes in the sample below the reference value in
each stratum (nc). The next step is to determine the
proportion of lakes in the sample below the refer-
ence value for each stratum (nc/n***:2/56 = 0.0357
and 20/46 = 0.4348). Next, multiply the proportion
of sample lakes below the reference value in the
stratum by the estimated number of lakes in the
stratum population (N), which results in Nc, the es-
timated number of lakes jn the population below
the reference value. The SNC for each stratum yields
the combined stratum Nc. The same answer can be
obtained by multiplying nc by W for each stratum
and summing the results.
The best estimate for the overall proportion of lakes in
the designated population below the reference value,
therefore, is 289.79/822.65 = 0.3523 (Table 4-2). If
the overall proportion of lakes below the reference
value were computed as 22/102 = 0.216 (nc/n*** for
the sum of nc and n*** for both strata), the answer
would be biased. For example, there is an estimated
total of 823 lakes in strata 2A1 and 2A2. Using the
value of 0.3523 as pc, the estimated number of lakes
with pH <6.0 would be 290. Using the pc value of
0.216 (based on the combined nc/n***), the estimated
number of lakes with pH <6.0 would be 112.
Therefore, the number of lakes estimated to have pH
<6.0 in both strata would be underestimated by 178
(290-112).
Table 4-2. Comparison of Weights in Two Strata, 2A1 and
2A2, Eastern Lake Survey- -Phase I
pH == 6.0
Stratum
N
n
W
2A1 170.13 56 3.038 2 0.0357 6.07
2A2 652.52 46 14.185 20 0.4348 283.72
Combined 822.65 102 22 0.3523 289.79
N = estimated number of lakes within an alkalinity map class
stratum.
n*** = number of lakes from which samples were obtained.
W = weighting or expansion factor.
nc = number of lakes in the probability sample with pH s 6.0,
the reference value.
Pc = estimated proportion of lakes in the sample or population
for a stratum or combined strata, respectively, which
have a pH < 6.0 (nc/n***).
Nc = estimated number of lakes in the population which have
a pHs 6.0, the reference value.
A less clear issue associated with the design con-
siderations and weighting is related to examining
relationships among variables. Unweighted analy-
ses, such as regressions or correlations, should be
used with caution. For the exploratory approach
used to examine a limited number of relationships
for this report, we have chosen to do unweighted
analyses. It should be noted, however, that the
parameter estimates (e.g., the regression intercept
or slope) do not necessarily represent those of the
population. It is of interest to explore the conform-
ity of relationships across strata and among vari-
ables of interest. Unless the relationships of inter-
est are independent of alkalinity map class (and any
factor associated with the alkalinity map class strata)
these estimates can be biased, just as unweighted
means or medians and total lake numbers can be.
4.1.2.2 Data Quality
An aspect of the analytical data quality that is useful
in interpreting the data is the system decision limit
(SDL, Section 3.5.2). Data below the SDL are not
confidently distinguishable from blanks, but still
have reported values which are meaningful in ex-
amining relationships among variables. The SDLs
are summarized for measured variables in Table
4-3. The SDL should be considered when selecting
reference values (Section 4.3.3.2) for comparing
populations. Reference values below the SDL
should not be selected if tests of significant differ-
ences are to be made. An exception is ANC with an
SDL of 6.9 jieq L~1. Because ANC can be negative,
only values within the range of -6.9 to 6.9 are not
distinguishable from blank values.
Substantial percentages of lakes in one or more
regions had concentrations of extractable Al, total
P, NH«+, NO/, total Al, Fe, Mn, Si02 and DIG below
the SDL (Table 4-3). These high percentages sug-
gest that many of the waters in the ELS-I are quite
dilute, as expected from the design of the Survey.
Of the primary variables, a large percentage of
lakes occurred below the SDL only for extractable
Al, which had an SDL of 8 (xg L~1. However, as with
many of the ELS-I variables, the high concentra-
tions of extractable aluminum are of primary inter-
est. For extractable aluminum, the SDL is well be-
low the concentration expected to affect fish, for
example (Section 5.2.4).
4.2 Description of the Target Population
4.2.1 Number of Lakes Sampled
A total of 2681 regular (probability sample) lakes
were selected from the map population. Of those,
805 were classified as non-target by examination of
large-scale maps, 151 were classified as non-target
33
-------
Table 4-3. Data Quality Summary, Eastern Lake Survey -Phase I
Variable
PRIMARY:
pH (closed)
ANC
Sulfate
Calcium
Al, ext.
DOC
SECONDARY:
Magnesium
Sodium
Potassium
Ammonium
Nitrate
Chloride
Fluoride
Al, total
Iron
Manganese
Silica
P, total
Conductance
DIC (open)
Color
Turbidity
System
Decision
Limit8
-------
when visited, and 113 were not visited (Section
2.2.4). Data from water samples collected from
1612 lakes were subsequently considered for use in
making population estimates. Twenty of these
lakes were separated from the remaining lakes in
the analyses because of their large size (Section
4.2.2).
Of the 199 special interest lakes selected (Section
2.2.5), 186 were sampled. The data collected from
special interest lakes are presented in Volumes II
and III, but are not discussed in this report. Because
these lakes were not part of the random selection
process, weighting factors do not apply in this case
(Section 4.1.2), and their representativeness with
respect to the chemical characteristics of the lake
populations as a whole is uncertain (Section 4.9).
4.2.2 Treatment of Large, Shallow and
Thermally-Stratified Lakes
Preliminary data analysis revealed that cumulative
areal distributions, G(x), were affected by 20 excep-
tionally large (>2000 ha) lakes. These lakes had lit-
tle effect on the cumulative frequency distributions,
F(x), but greatly influenced the precision of G(x)
(Table 4-4). For this reason, this subpopulation of
lakes was excluded from this report and only those
lakes <2000 ha (1592) are included in the analyses.
It is estimated that there are 197 such large lakes in
the study regions (61 in Region 1, 73 in Region 2,
and 63 in Subregions 3A and 3B). These large lakes
were not eliminated from the target population but
rather, through their exclusion in the analyses, de-
fine a new subpopulation of interest. Estimates of
numbers and characteristics of subpopulations of
lakes >2000 ha are available, but precision is poor.
An objective of sampling lakes in the fall was to
obtain one sample at a standard depth (1.5 m) that
would best represent the entire lake chemistry. Of
the 1612 probability sample lakes, 17.4 percent
were sampled at a depth less than the originally
prescribed depth of 1.5 m. This adjustment to the
protocol for shallow lakes was required to minimize
contamination of the sample from sediment.
Of the 1612 probability sample lakes sampled, three
percent were weakly stratified and two percent
strongly stratified (Section 2.6.2). Because hypo-
limnetic waters may contain higher concentrations
of ANC, the surface waters of stratified lakes could
have exhibited somewhat higher ANC values fol-
lowing mixing than were measured while the lakes
were stratified. Thus, the question arose as to how
representative of the entire water mass a single
sample might be.
The possible bias introduced from including strati-
fied and shallow lakes collectively in the analyses
was a concern. However, because little difference
was observed between population proportions for
ANC based on including or excluding stratified or
shallow lakes (Section 4.5.1.6), population esti-
mates for this report are based on all lakes sampled
for the subpopulation of lakes restricted by size (all
lakes in the target population <2000 ha).
4.2.3 Target Population and Population
Estimates
Table 4-5 provides the basic structural computa-
tions for the sample and target populations dis-
Tab!e4-4.
Subregional and Regional Summaries of Estimated Target Population Size (N), Estimated Target Population Area (A),
and the Standard Errors (SE) of These Estimates, Eastern Lake Survey -Phase I
REGION
1
SUBREGION 2A
28
2C
2D
REGION 2
SUBREGION 3A
3B
All Lakes*
Lakes <2000 ha**
SE(N)
A
SE(A)
SE(NS)
7157
1499
1050
1511
4515
8575
286
2138
163.9
72.4
72.5
45.7
293.9
314.6
20.7
214.5
1093589
1484385
34025
256444
226896
2001749
344465
312936
458727
1214881
10854
107928
58202
1221102
138586
232528
7096
1457
1050
1480
4515
8502
258
2098
165.3
74.2
72.5
48.8
293.9
315.5
20.5
212.6
427865
142981
34025
97556
226896
501458
24272
66169
SE(AS)
SUBREGION 1A
1B
1C
1D
1E
1290
1506
1494
1325
1542
47.6
90.9
57.0
93.6
65.7
118777
478659
101578
95940
298636
22186
443078
29948
54885
98546
1290
1479
1483
1318
1526
47.6
92.9
57.5
93.7
66.0
118777
26872
72412
36403
173400
22186
2913
12013
4381
25725
36414
37582
10854
21357
58202
73307
4795
10158
•Including lakes >2000 ha.
"•Subscript s denotes a subpopulation, in this case those lakes in the target population =£2000 ha.
35
-------
Table 4-5. Description of Sample and Target Population
(Stratum Specific), Eastern Lake Survey -Phase I
SE(N) A SE(A)
STR N* n* n*** W
N
1A1 711
1A2 542
1A3 431
1B1 208
1B2 96
1B3 1682
1C1 631
1C2 752
1C3 650
1D1 443
1D2 656
1D3 1568
75 57 9.633 549.08 33.08
65 51 8.338 425.24 26.13
68 47 6.719 315.79 22.14
78094 20460
21460 5690
19223 6420
70 49 3.192 156.41 9.29 3851 544
70 48 1.477 ^0.90 3.00 2046 233
68 47 27.209 1278.82 90.37 472762 443077
88 63 7.822 492.79 27.31
70 54 10.743 580.12 36.20
74 47 8.953 420.79 34.59
70 47 6.572 308.88 23.00
95 43 6.905 296.92 31.14
93 37 19.426 718.76 85.22
26943 9848
53804 27942
20831 4376
67439 54729
8977 1695
19523 3784
1E1 1038 130 89 8.070 718.23 39.71 89853 26873
1E2 606 74 48 8.344 400.51 31.80 138490 93036
1E3 744 72 41 10.333 423.65 41.55 70292 18259
2A1 176 60 56 3.038 170.13 3.43 17419 3382
2A2 778 62 46 14.185 652.51 37.60 105222 55019
2A3 1178 85 48 14.098 676.70 61.79 1361744 1213630
2B1 118 74 41 1.878 77.00 4.89 893 90
2B2 250 100 57 2.579 147.00 9.83 2776 500
2B3 1330 80 48 17.208 825.98 71.69 30357 10842
2C1 464 60 50 8.340 417.00 17.77 8044 1472
2C2 348 60 56 6.007 336.39 7.56 10432 1382
2C3 895 60 49 15.459 757.49 41.38 237968 107910
2D1 97 90 40 1.536 61.44 3.50 1071 94
2D2 699 85 53 9.251 490.30 34.99 9027 2103
203 5351 70 48 82.558 3962.78 291.82 216797 58164
3A1 19 19 11 1.071 11.78 0.58 874 96
3A2 76 60 47 1.343 63.12 1.93 24432 3914
3A3 443 100 44 4.792 210.85 20.58 319159 138531
3B1 1608 140 52 13.127 682.60 69.30 21763 3917
3B2 113 113 62 1.000 62.00 0.00 18705 0
3B3 6332 181 36 38.705 1393.38 203.05 272468 232495
STR = Stratum W = Expansion factor
N* = Frame population N = Estimated target popula-
n* = Number of lakes in tion size
the probability SE(ft) = Standard error of N
sample A = Estimated area of target
n*»* = Number of lakes _ population
sampled SE(A) = Standard error of A
cussed in this report, by strata. All estimates are
made with the equations in Section 2.3.2. The
weights provided in Table 4-5 are appropriate for
any weighted analyses that might be necessary
(Section 4.1.2.1).
4.3 Descriptive Statistics and Cumulative
Distributions
The primary objectives of the ELS-I were to charac-
terize the chemistry of lakes in potentially sensitive
regions of the United States, to estimate the num-
bers of low ANC and low pH lakes, to identify where
those lakes are, and to describe their characteris-
tics. The Survey addresses these objectives by its
stratified design which permits description of the
populations of lakes in each of the 33 strata, repre-
senting the three alkalinity map classes within each
of the 11 chosen subregions. These strata represent
the basic analytic units, but as noted earlier, analy-
ses may also be performed on any combinations of
strata that have meaning to the investigator (Sec-
tion 2.2.1) within regions, subregions, or subpopu-
lations (for example, lakes <2000 ha).
4.3.1 Population Distributions
Each population to be described was summarized
on a single page of output for each variable. The
cumulative distributions, F(x) and G(x), supple-
mented by descriptive statistics, were chosen for
this purpose and serve as the primary data outputs.
Figures 4-1 to 4-8 for ANC and field laboratory pH
are examples. Volume II includes these outputs and
those for other variables and subpopulations. Ta-
bles of summary statistics are generated from
these outputs. Direct inspection of the differences
among the elements or strata of the design struc-
ture can be made by using a combination of the
distribution plots and tabular information provided
on these data outputs.
By presenting the data as cumulative distributions,
the utility and flexibility of ELS-I are enhanced.
Specifically, low pH or low ANC, for example, are
not unequivocally or universally defined. There-
fore, several reference values for each of the pri-
mary variables were chosen, primarily to facilitate
comparisons among strata, subregions, regions, or
subpopulations. However, the strength of the de-
signed population description output, as well as of
the Survey design itself, is that any reference value
can be evaluated, providing flexibility in interpreta-
tion of the output. As the understanding of acidifi-
cation processes improves, the reference values
can be changed, ensuring the utility of the ELS-I
data base in the future. The considerations noted in
Section 4.1.2.2, however, are important in selecting
additional reference values and Section 4.3.2 pro-
vides examples of the interpretive flexibility in
using the outputs.
4.3.2 Definitions of the Descriptive Statistics: In-
terpretation of FM and GM Data Output
Each population description (Figures 4-1 through
4-8) is identified by the variable (chemical parame-
ter) and by the subset of lakes constituting the pop-
ulation being examined. A subset may represent all
lakes in the Survey unit, as for example, a single
stratum (1A1), a subregion (1A), a region (1), or a
subpopulation.
36
-------
Figure 4-1.
F(x) and G(x) distributions of ANC (//eq L~1) for the target population of lakes (<2000 ha) sampled in Region 1
(Northeast), Eastern Lake Survey-Phase I.
16 Feb. 1986
Variable: ANC as /veq L"1 for lakes <2000 ha
Xci: 0.0 XC2: 50.0 XC3:200.0
Population Size (N): 7096 SE(N): 165.3
Lake Area (A): 427864 SE(A): 36414
Sample Size: 763
Region 1
Subregion
Class
1.0
0.8-
0.6-
0.4-
0.2-
0.0
-100.0' 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0, 900.0 1000.0
Min:-45.60 Qi: 51.63 Q2: 118.36 Q3: 199.64 Q4: 399.94 Max: 4046.60
Median: 158.11
Mean: 268.08
Std. Dev.: 411.69
Proportions and Numbers Below the Value of Xc
Pci: 0.046 Nci: 326 Ncui: 422
r»__* A 1 QO M _• 1 QO1 Kl -• 1 KQR
Pea: 0.192
PCS: 0.600
NC2: 1364
NC3: 4258
: 1536
Nous: 4513
1.0
0.8-
0.6-
0.4-
0.2-
0.0-
-100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
Min:-45.60 Qi: 68.21 Q2: 113.61 Q3: 172.01 Q4: 349.45 Max: 4046.60 :
Median: 131.94
Proportions and Numbers Below the Value of Xc
Qci: 0.018 AC1: 7756 Acui: 11567
9c2: 0.102 AC2: 43839 Aeu2: 56963
Qcs: 0.672 AC3: 287422 Acu3: 338943
37
-------
Figure 4-2.
F(x) and G(x) distributions of ANC (jueq L"1) for the target population of lakes (<2000 ha) sampled in Region 2 (Upper
Midwest), Eastern Lake Survey-Phase I.
16 Feb. 1986
Variable: ANC as fieq/L for Lakes <2000 ha
Xci: 0.0 XC2: 50.0 XC3: 200.0
Population Size (N): 8501 SE(N): 315.3
Lake Area (A): 501547 SE(A): 73306
Sample Size: 587
Region 2
Subregion
Class
1.0
0.8-
0.6-
0.4-
0.2-
0.0
-100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
Min: - 48.60 QI: 77.82 Q2: 183.94 Q3: 642.97 Cu: 1415.51 Max: 4002.00
Median: 359.54
Mean: 755.98
Std. Dev.: 881.14
Proportions and Numbers Below the Value of Xc
pci:0.017 Nci: 148 Ncui: 209
DCS: 0.154 N,,: 1312 N,.,,,: 1605
Pcz:0.154
Pea: 0.414
NC2: 1312
NC3: 3518
NCU2: 1605
Nous: 3982
1.0-
0.8-
.0.6-
C3
0.4-
0.2i
0.0
-100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
Min: - 48.60 Qi: 265.36 Q2: 592.95 Q3: 936.58 Q4: 1860.05 Max: 4002.00
Median: 680.68
Proportions and Numbers Below the Value of Xc
gci: 0.003 Aci: 1513 Acui: 2191
n-v r>D33 A-,: 16737 A.,,,: 20609
gc2: 0.033
Qca: 0.152
Ac2: 16737
AC3: 76374
ACU2: 20609
Aou3. 88914
38
-------
Figure 4-3. F(x) and G(x) distributions of ANC (jueq L~1) for the target population of lakes (<2000 ha) sampled in Subregion 3A
{Southern Blue Ridge), Eastern Lake Survey-Phase I.
16 Feb. 1986
Variable: ANC as fjeq/L for Lakes <2000 ha
Xcr. 0.0 XC2: 50.0 XC3: 200.0
Population Size (N): 258 SE(N): 20.5
Lake Area (A): 24272 SE(A): 4795
Sample Size: 94
Region 3
Subregion A
Class
0.2-
0.0
-100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
Min: 20.20 Qt: 149.15 Q2: 225.74 Q3: 265.80 Cu: 421.36 Max: 1863.55
Median: 250.24
Mean: 303.36
Std. Dev.: 216.83
Proportions and Numbers Below the Value of X0
Pc1: 0.000 Nci: 0 Ncui: **
Pc2: 0.014 NC2: 3 Ncu2: 5
Po3: 0.343 NC3: 89 NCU3: 108
0.2-
0.0
-100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
\
Mm: 20.20 Qi. 113.16 Q2: 151.44 Q3: 258.09 Q4: 316.83 Max: 1863.55
Median: 185.45
Proportions and Numbers Below the Value of Xc
g0i: 0.000 A0i: 0 Acui: **
Qci 0.003 AC2: 79 ACU2: 108
gc3: 0.546 AC3: 13258 ACU3: 16436
39
-------
Figure 4-4. F(x) and G(x) distributions of ANC (yueq L~1) for the target population of lakes (<2000 ha) sampled in Subregion 3B
(Florida). Eastern Lake Survey-Phase I.
16 Feb. 1986
Variable: ANC as fjeq/L for Lakes <2000 ha
Xoi: 0.0 X02: 50.0 Xc3: 200.0
Population Size (N): 2098 SE(N): 212.6 Sample Size: 148
Lake Area (A): 66169 SE(A): 10158
Region 3
Subregion B
Class
1.0-
0.8-
0.6-
0.4-
0.2-
0.0
-100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0
Min: -209.12 Qi: 7.48 Q2: 68.56 Q3: 275.05 CU: 713.05 Max: 3639.76
Median: 83.51
Mean: 535.18
Std. Dev.: 871.42
Proportions and Numbers Below the Value of Xc
Pc1: 0.220 Nc,: 463 Noui: 615
Pc2: 0.353 NC2: 742 Ncua: 942
Pc3: 0.551 NC3: 1156 NCU3: 1413
1.0
0.8-
0.6-
0.4-
0.2-
0.0
-100.0. 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1000.0:
Min: -209.12 Q,: -2.06 Q2: 46.72 Q3: 142.97 Q4: 473.83 Max: 3639.76
Median: 72.21
Proportions and Numbers Below the Value of Xc
gc1: 0.202 Acl: 13348 Acui: 18790
gcl- 0.435 A02: 28778 ACU2: 414020
gc3- 0.641 AC3: 42434 Acu3: 55674
40
-------
Figure 4-5. F(x) and G(x) distributions of pH (closed system) for the target population of lakes (<2000 ha) sampled in Region 1
(Northeast), Eastern Lake Survey -Phase I.
16 Feb. 1986
Variable: pH for Lakes <2000 ha
Xci: 5.0 XC2: 5.5 XC3: 6.0
Population Size (N): 7096 SE(N): 165.3
Lake Area (A): 427864 SE(A): 36414
Sample Size: 763
Region 1
Subregion
Class
1.0
0.8-
0.6-
0.4-1
0.2
0.0
4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
Min: 4.32 Qi: 6.33 Qj,: 6.74 Q3: 6.97 Q4: 7.26 Max: 9.36
Median: 6.87
Mean: 6.76
Std. Dev.: 0.74
Proportions and Numbers Below the Value of Xc
Pc1: 0.034 Nci: 240 Ncul: 315
pcz: 0.086 Net 613 Ncu2: 737
Pc3: 0.129 NC3: 916 Ncu3: 1056
1.0
0.8-
0.6-
0.4-
0.2-
0.0
4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5
Min: 4.32 Qi: 6.58 Q2: 7.09 Q3: 7.09 Q«: 7.37
8.0 8.5 9.0
Max: 9.36
Median: 6.94
Proportions and Numbers Below the Value of X0
gci: 0.012 Aci: 5059 ACU1: 7570
go* 0.043 AC2: 18195 ACU2: 26393
flea: 0.060 Ac3: 25800 Acu3. 34478
41
-------
Figure 4-6. F(x) and G(x) distributions of pH (closed system) for the target population of lakes (<2000 ha) sampled in Region 2
(Upper Midwest), Eastern Lake Survey-Phase I.
16 Feb. 1986
Variable: pH for Lakes <2000 ha
Xc,: 5.0 XC2: 5.5 Xo3: 6.0
Population Size (N): 8501 SE(N): 315.5
Lake Area (A): 501458 SE(A): 73306
Sample Size: 587
Region 2
Subregion
Class
Max: 8.69
Median: 7.09
Mean: 7.09
Std. Dev.: 0.82
Proportions and Numbers Below the Value of Xe
Pci: 0.015 Nci: 130 Ncul: 189
Pea: 0.036 Ne2: 309 Ncu2: 390
PCS: 0.096 NC3: 818 Ncu3: 1037
o
0.0
4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5
Min: 4.43 Qi: 7.04 Q2: 7.35 Q3: 7.63 Q4: 7.98
8.0 8.5 9.0
Max: 8.69 '-.
Median: 7.44
Proportions and Numbers Below the Value of Xc
8ci: 0.002 Act: 1064 Acuv 1553
gc2: 0.006 Ae2: 3055 Acu2: 3942
gc3: 0.021 AC3: 10399 Acu3: 13686
42
-------
Figure 4-7. F(x) and G(x) distributions of pH (closed system) forth* target population of lakes (<2000 ha) sampled In Subregion
3A (Southerm Blue Ridge), Eastern Lake Survey-Phase I.
16 Feb. 1986
Variable: pH for Lakes <2000 ha
Xci: 5.0 Xe* 5.5 Xc* 6.0
Population Size (N): 258 SE(N): 20.5
Lake Area (A): 24272 SE(A): 4797
Sample Size: 94
1.0
0.8-1
0.6-
F(x)
0.4-
0.2-
0.0
Median: 6.98
Mean: 7.01
Std. Dev.: 0.37
Proportions and Numbers Below the Value of Xc
Pei: 0.000 NCI: 0 Neu,: •»
Pet: 0.000 Ne2: 0 Ncu2: •*
p.* 0.004 No: 1 NCU3: 2
1.0
0.8-
0.6-
G(x)
0.4-
0.2-1
0.0
4.0 4.5 5.0
Min: 5.90 Q,: 6.76
5.S 6.0 6.5
Q2: 6.82 Qa: 6.93
4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
Min: 5.90 Q,: 6.73 Q2: 6.92 Q,: 7.08 Q4: 7.29 Max: 8.27
7.0 7.5 8.0 8.5
Q«. 7.11 Max: 8.27
9.0
Median: 6.92
••Not Defined
Proportions and Numbers Below the Value of Xc
ge1: 0.000 Aei: 0 Aeui: **
get: 0.000 Ae2: 0 ACU2: **
gea: 0.001 Acs: 13 Aeua: 18
43
-------
Figure 4-8. F(x) and G(x) distributions of pH (closed system) for the target population of lakes (<2000 ha) sampled in Subregion
3B (Florida), Eastern Lake Survey-Phase I.
Variable: pH for Lakes <2000 ha
Xc,: 5.0 XC2: 5.5 XC3: 6.0
Population Size (N): 2098 SE(N): 212.6
Lake Area (A): 66169 SE(A): 10158
16 Feb. 1986
Sample Size: 148
Region 3
Subregion B
Class
1.0
0.8-
0.6-
F(x)
0.4-
0.2-
0.0
Median: 6.56
Mean: 6.45
Std. Dev.: 1.03
Proportions and Numbers Below the Value of Xc
Pcv 0.124 NC1: 259 Noul: 385
Pea: 0.206 NC2: 433 N«,,»: 574
Pea: 0.206
Pci 0.327
NC2: 433
3: 687
: 574
NCU3: 878
1.0
0.8-
0.6-
G(x)
0.4-
0.2-
0.0
4.0 4.5 5.0
Min: 3.81 Q,: 5.36
5.5 6.0 6.5
Q2: 6.41 Q3: 6.81
7.0 7.5 8.0
Q«: 7.47 Max: 8.96
Median: 6.76
Proportions and Numbers Below the Value of Xc
gci: 0.120 Aci: 7936 Acui: 12666
Qca: 0.235 Acz: 15553 ACu2: 21925
9c3'. 0.327 AC3: 21635 Acu3: 28889
4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
Min: 3.81 Q,: 5.38 Q2: 6.37 Q3: 6.84 Q4: 7.41 Max: 8.96
8.5
9.0
44
-------
The basic population estimates, N and A, and the
estimated standard errors of these two estimates
are presented (Figures 4-1 through 4-8) in the head-
ings of the distribution curve output. N is defined as
the estimated total number of lakes in the popula-
tion being described, and A is the estimated sum of
the total lake area, over all lakes in that population
(denoted only as N and A on the output). The indi-
cated sample size (equivalent to n***. Section
2.3.2) is the number of sampled lakes in the target
population being described. Identity of reference
values (Xc1, Xc2, X^) chosen for comparative analy-
ses completes the heading material on each page
of output.
4.3.2.1 Distributions
The distribution curves, F(x) and G(x), are similar in
interpretation, representing estimated frequency
and areal distributions, respectively. For any refer-
ence value for variable X, F(x) is interpreted as the
estimated proportion of lakes having a value of X
=sx. F(x) is calculated as the ratio of Nx, the esti-
mated number of lakes having a value of X
-------
Ncu = the upper confidence bound on
number of lakes,
gc = the estimated proportion of lake
area,
Ac = the estimated area of lakes, and
Acu = the upper confidence bound on
area of lakes having the value of
X < Xe (or > Xc).
• For single strata only, an upper confidence limit
also is provided for pc (Volume II).
4.3.3 Comparisons of Distributions
One of the most informative ways to examine dif-
ferences among population distributions is by su-
perimposing F(x) plots. Examples of how such
plots, weighted for the differences in lake popula-
tion numbers within/among strata, can be inter-
preted are presented below for pH and ANC. With
regard to pH (Figure 4-9), Subregion 1A is distinct
from the other subregions in Region 1, having
many more lakes in the lower tail of the distribu-
tion. In Region 2, there is considerably more varia-
tion among the subregions than in Region 1, with
all subregions but 2A having substantially more
variation from Region 1. Subregion 2A, however,
has a very different distribution than others in Re-
gion 2, with greatly lower within-population varia-
tion. Region 3 should not be considered as a whole
because the two subregions are markedly different.
Subregion 3A has very little within-population vari-
ation, while the pH distribution in Subregion 3B
showed considerable variation.
The plots of ANC distributions (Figure 4-10) show,
as expected, similarities to the pH plots. In Re-
gion 1, a family of curves is seen for 1A-1D, with 1A
shifted to the left. Acid neutralizing capacity is
higher in Subregion 1B. Subregions 1A and 1B es-
sentially form the envelope of the distributions of
the two variables, pH and ANC. Subregion 1E has
few lakes of low pH, but many lakes of low ANC. In
Region 2, the pattern of ANC in Subregions 2A, 2B,
and 2D is consistent with that of the pH curves.
Many sites have low ANC in Subregion 2C. A few
acidic lakes (ANC <0 jieq L~1) occur in Subregion
2B, but most lakes in the subregion have higher
ANC than other parts of the Upper Midwest. The
ANC distribution for 3B (Florida) was very similar to
that for 2C (Northcentral Wisconsin). In both cases
there were many low ANC systems, largely seep-
age lakes (Section 4.4), and then a change in the
shape of the curve due to higher ANC in other lake
types. A straightline increase in F(x) was observed
in both pH distributions.
Other similarities in the F(x) for ANC among subre-
gions were observed for Subregions 1 E, 2A and 3A.
The shape of the curve for 1A was also related, but
apparently shifted as a result of more low ANC
lakes.
4.3.4 Interpretation of Alkalinity Map Classes
The third level of stratification in the ELS-I design
(Section 2.2.1) was alkalinity map class. Compari-
sons among alkalinity map classes provide evi-
dence as to the effectiveness of this level of stratifi-
cation.
The maps showing the alkalinity map classes are
provided in Figures 2-2 through 2-4. Because of the
map scale, there will be a diversity of ANC values
for lakes found within any one alkalinity map class.
If the alkalinity map classes are useful as stratifica-
tion factors, a random sample of lakes within a map
class will yield more lakes of the designated class
than any other class.
In Regions 1 and 2, the largest percentage of lakes
that had ANC values <100 n-eq L~1 occurred in map
class 1 and the lowest percentage in map class 3
(Table 4-6). Subregion 3A had few lakes (12) in map
class 1 but the results were as expected. Subregion
3B had many class 1 and 3 lakes, but few class 2
lakes. However, the largest percentage of lakes with
ANC <100 neq L"1 occurred in map class 1, as an-
ticipated.
Results similar to those expected for alkalinity map
class 3 were observed (Table 4-6). Alkalinity map
class 2 was apparently less effective than map
classes 1 and 3 as a stratification factor. The original
maps were developed from data available at the
time. Most of these data were not obtained in the
fall nor were they obtained within the same year for
all systems. Considering seasonal and yearly ANC
variability, it was therefore not surprising to ob-
serve many lakes with ANC <100 (xeq L~1 or
>200 jxeq L~1 within map class 2. Map class 2 was,
however, an effective stratification factor in the
Southern Blue Ridge. In Florida (3B) very few lakes
with observed ANC from 100 to 200 n,eq L"1 oc-
curred in map class 2; thus, there was poor agree-
ment between the observed percentage of lakes
with ANC in this range and the expected percent-
age based on map class 2.
In conclusion, the field data suggest that the strati-
fication was effective, but the effectiveness varied
by region and map class, and within subregional
strata (not shown). This is consistent with the in-
tended use of the alkalinity maps (i.e., they are a
meaningful representation of spatial alkalinity pat-
terns). However, even though the stratification was
effective, the population of lakes in each map class
within a region will not be entirely in that alkalinity
class. Each map class is a mixture of alkalinity
classes. For example, in Region 2, there are nearly
46
-------
Figure 4-9. Comparisons among subregions of cumulative frequency distributions [F(xJ] for pH (closed system). Eastern Lake
Survey-Phase I.
F(x)
I I I I I I
__ Region 1
0.2 h—
0.0
1.0
0.8
0.6
F(x)
0.4
0.2
0.0
1.0
0.8
0.6
F(x)
0.4
0.2
0.0
I I I I I
I I
Region 2
l^III
I I I I
I I I I I I
__ Region 3
.\>'' I
I I
4.0 4.5 5.0
5.5 6.0 6.5 7.0 7.5
pH (Closed System)
I I I I
8.0 8.5 9.0
47
-------
Figure 4-10. Comparisons among subregions of cumulative frequency distributions [F(x|] for ANC (fjoq L~1) Eastern Lake
Survey-Phase).
F(x)
1.0
0.8
0.6
:)
0.4
0.2
0.0
1.0
0.8
0.6
Region 1
F(x)
0.4
0.2
0.0
F(x)
200
400
ANC (/aeq L'1)
600
800
1000
48
-------
Table 4-6. Composition of the Alkalinity Map Classes, in
Numbers and Percentage of Lakes Having Meas-
ured ANC (Aieq L~1) in those same Classes:
1 }<100,2) 100-200,3) >200 for Lakes <2000
ha. Eastern Lake Survey-Phase I*
Estimated Percentage
Region 1
Map Class 1
Map Class 2
Map Class 3
Region 2
Map Class 1
Map Class 2
Map Class 3
Subregion 3A
Map Class 1
Map Class 2
Map Class 3
Subregion 3B
Map Class 1
Map Class 2
Map Class 3
1
55.1
46.3
14.4
73.4
40.0
16.2
45.4
22.7
5.1
73.1
52.5
40.0
2
33.1
27.4
18.1
14.8
36.9
10.5
27.3
47.7
15.4
5.8
8.2
2.9
3
11.8
26.3
67.5
11.8
23.1
73.3
27.3
29.6
79.5
21.1
39.3
57.1
k0lll 1 IUIWU
Number
2,211
1,755
3,130
726
1,598
6,177
12
59
187
683
61
1,354
*This comparison can also be made at the stratum level within
subregions to examine the effectiveness of the alkalinity map
class stratification. Although there is considerable variability
among strata within subregions, and thus the effectiveness of
stratification also varies at this scale, the above presentation
suggests that within the regions as a whole the alkalinity map
class stratification was appropriate.
as many observed class 1 lakes in alkalinity map
class 3 as in classes 1 and 2 combined [(0.162
x 6177 = 1001) a (0.734 x 726 = 533) + (0.400
x 1598 = 639}].
4.4 Physical Characteristics of Regional
Lake Populations3
4.4.1 Northeast
Northeastern watersheds were generally larger
than those in the Upper Midwest and Florida (3B),
but smaller than those in the Southern Blue Ridge
(3A; Table 4-7). Maine (1E) had the largest median
watershed area in the Northeast (472 ha), while the
Poconos/Catskills (1B) had the smallest (169 ha).
Maine (1E) had the largest median lake area (29.1
ha; the Adirondacks (1A) had the second largest
(20.9 ha). Lakes in the Poconos/Catskills (1B) were
uniformly small (median = 12.7 ha, fourth quintile,
Cu = 22.6 ha). Site depths in the Poconos/Catskills
(1 B) and Southern New England (1D) were among
the shallowest of all subregions (median site depth =
3AII references to medians, quintiles, percentages, lake numbers, and
areas are estimates of population values but are not always noted as
estimates in the text. Information on upper confidence limits is provided
in the tables but is not discussed in the text. The values for pc (propor-
tions) in the tables can be converted to percentages by multiplying pc by
100.
3.2 m and 2.9 m, respectively)(Table4-7). Drainage
lakes predominated in the Northeast (71% overall;
Table 4-8); in the Poconos/Catskills (1 B), reservoirs
were common (42%); reservoirs were also common
in Southern New England (10; 29%), as were
seepage lakes (13%). Of the northeastern (Region 1)
lakes, those in the Adirondacks (1A) and the
Poconos/Catskills(1B) were at the highest elevation
(Table 4-7). Roads and dwellings were the pre-
dominant watershed disturbances in all subregions,
although logging was also commonly observed in
Maine (1E).
4.4.2 Upper Midwest
Median watershed areas in all subregions of the
Upper Midwest were less than those in all subre-
gions of the Northeast except the Poconos/Catskills
(1B, Table 4-7). Median watershed areas in the Up-
per Peninsula of Michigan (2B) and Northcentral
Wisconsin (2C) were among the smallest of any of
the subregions (115 and 112 ha, respectively).
Lakes in the Upper Peninsula of Michigan (2B) were
also among the smallest (median area = 11.1 ha)
and shallowest (median depth = 2.9 m) of any sub-
region (Table 4-7). Drainage lakes (50%) and seep-
age lakes (43%) were the most common lake types
in the Upper Midwest (Table 4-8). Seepage lakes
were predominant in Northcentral Wisconsin (2C;
59%) and the Upper Great Lakes Area (2D; 47%),
although drainage lakes were more common in the
Upper Peninsula of Michigan (2B; 51%) and in
Northeastern Minnesota (2A; 74%). Roads and
dwellings were the predominant watershed dis-
turbances in the Upper Midwest. Roads and
dwellings were half as frequent in Northeastern
Minnesota (2A) and the Upper Peninsula of Michi-
gan (2B) as in the remainder of the region.
4.4.3 Southeast
Large differences were found between the popula-
tions of lakes in the two subregions in the South-
east (Region 3). Consequently, population esti-
mates for these subregions have not been
combined to produce regional estimates.
4.4.3.1 Southern Blue Ridge—Subregion 3A
Lakes in the Southern Blue Ridge had the largest
median watershed area (682 ha), were among the
deepest (Q4 = 12.5 m), and were the smallest (me-
dian area = 10.8 ha) of any subregion (Table 4-7).
Ninety percent of the lakes in Subregion 3A were
reservoirs (Table 4-8). Lakes in the Southern Blue
Ridge were generally higher in elevation (me-
dian = 265 m) and had a greater frequency of
watershed disturbances than those in any other
subregion. The primary disturbances were roads
and dwellings, although logging and other land
uses were also important.
49
-------
Table 4-7. Physical Lake Characteristics: Medians (M) and First and Fourth Quintiles (Qi and QH\, Eastern Lake Survey -Phase 1
Lake Watershed Lake Site
Elevation (rn) Area (ha) Area (ha) Depth (m)
SUBREGION 1A
1B
1C
1D
1E
REGION 1
SUBREGION 2A
2B
2C
2D
REGION 2
SUBREGION 3A
3B
QI
339
232
150
34
69
110
426
224
480
331
339
221
15
Table 4-8. Population
SUBREGION 1A
1B
1C
ID
1E
REGION 1
SUBREGION 2A
2B
2C
2D
REGION 2
SUBREGION 3A
3B
M
486
361
295
171
136
307
458
267
494
387
408
265
21
Estimates of
Reservoir
100 (8)
622 (42)
70 (5)
381 (29)
43 (3)
1216 (17)
31 (2)
52 (5)
15 (1)
84 (2)
182 (2)
232 (90)
0 (0)
Q4
566
479
408
601
316
491
534
451
510
416
487
454
31
Lake Type:
QI
114
72
111
65
123
90
69
46
47
59
56
182
35
M
265
169
353
237
472
271
201
115
112
181
177
682
115
: Number of Lakes
Drainage
991
799
1206
690
1386
5072
1074
539
529
2080
4222
18
452
(77)
(54)
(81)
(52)
(91)
(71)
(74)
(51)
(36)
(46)
(50)
(7)
(21)
0,
1961
430
1564
558
3187
1289
1681
659
575
1578
1023
5836
352
,* Eastern
QI
8.4
8.2
7.4
7.5
8.7
8.0
8.7
5.1
6.4
8.0
7.6
5.2
6.5
M
20.9
12.7
14.3
15.0
29.1
16.7
17.0
11.1
15.6
16.2
14.8
10.8
17.3
Lake Survey -
Closed
89
0
105
80
56
330
119
64
60
208
451
7
265
(7)
(0)
(7)
(6)
(4)
(5)
(8)
(6)
(4)
(5)
(5)
(3)
(13)
04
84.7
22.6
67.8
44.8
140.3
62.7
70.0
22.2
77.6
37.4
46.1
64.3
44.1
Phase 1
Seepage
110 (8)
58 (4)
102 (7)
167 (13)
41 (3)
478 (7)
232 (16)
396 (38)
876 (59)
2142 (47)
3646 (43)
0 (0)
1380 (66)
QI
2.2
1.5
1.8
1.2
1.8
1.6
2.2
1.0
3.9
2.2
2.2
2.3
1.5
M Q4
5.6 10.1
3.2 7.3
4.5 8.3
2.9 5.9
4.9 10.6
4.2 8.3
5.3 11.6
2.9 7.6
6.0 10.0
5.9 11.3
5.6 10.9
4.8 12.5
2.7 5.2
Estimated**
Total Number
of Lakes
1290
1479
1483
1318
1526
7096
1457
1050
1480
4515
8501
258
2098
*Percentages are shown in parentheses.
**May differ slightly from row total due to rounding error.
4.4.3.2 Florida—Subregion 3B
Lakes in Florida (Subregion 3B) had the lowest me-
dian elevation (21 m) of any subregion. These lakes
had small watersheds (Q4 = 352 ha) in which roads
and dwellings were commonly observed. They
were among the shallowest (median site
depth = 2.7 m) of any subregion (Table 4-7). Seep-
age lakes were most common (66%), followed by
drainage lakes (21%) and closed lakes (13%). No
reservoirs were sampled in Florida (Table 4-8).
4.5 Regional, Subregional and State
Population Estimates: ANC and pH
4.5.1 Acid Neutralizing Capacity
4.5.1.1 Reference Values
Three reference values for ANC have been given: 0,
50, and 200 jneq L~1. These reference values were
50
selected for the following reasons. A sample with
an ANC of 0.0 fxeq Lr1 or less is acidic by definition.
A reference value for ANC of 50 ixeq L~1 was chosen
because lakes with this ANC value, while having
some acid neutralizing capacity, may experience
important decreases in ANC as a result of episodic
events (i.e., snowmelt or heavy rains). A concentra-
tion of ANC £200 |xeq L~1 has been used frequently
as a value separating "sensitive" lakes from other
lakes (Section 5.1.3). The data shown in the tables
for ANC <50 jjieq Lr1 and ANC <200 |xeq Lr1 are
cumulative. Thus, ANC <50 jxeq L~1 includes lakes
with ANC <0 jxeq Lr1; ANC <200 n,eq L~1 includes
lakes with ANC <0 |xeq L~1 and <50 (xeq L~1.
4.5.1.2 Northeast
In the Northeast, acidic lakes (ANC <0 jxeq Lr1)
were concentrated in the Adirondacks, in the
-------
Poconos, on Cape Cod, along the Rhode Island/
Connecticut border, in northcentral Massachusetts/
southwestern New Hampshire, and in central
Maine (Figure 4-11).
Within the Northeast, the highest number (138) and
percentage (11%) of acidic lakes were estimated to
have occurred in the Adirondacks (1A; Table 4-9).
The largest area (2937 ha) and highest areal per-
centage (11%) of acidic lakes were in the Poconos/
Catskills (1B). Maine (1E) was estimated to have
few acidic lakes. The highest number (459) and per-
centage (36%) and the greatest area (14,504 ha) of
lakes with ANC <50 jieq L~1 also was estimated for
the Adirondacks (1 A; Table 4-10). The highest areal
percentage (21%) of lakes with ANC ^50 jieq L~1
occurred in the Poconos/Catskills (1B). Maine (1E)
was estimated to have the fewest (165) lakes with
ANC <50 //eq L~1, but these represented a large area
(11,428 ha). In Region 1, except in the Poconos/
Catskills (1B), the majority of lakes was estimated
to have ANC <200 p,eq L~1 (Table 4-1 1 ). Sixty-seven
percent of the lake area in the Northeast had ANC
<200
4.5.1.3 Upper Midwest
Within the Upper Midwest, most of the acidic lakes
occurred in the Upper Peninsula of Michigan (2B;
102) and in Northcentral Wisconsin (2C, 45; Table
4-9). No acidic lakes were sampled in Northeastern
Minnesota (2A) or in the Upper Great Lakes Area
(2D). In the Upper Midwest, lakes with ANC
:£50 (xeq L~1 were located in Northcentral Wiscon-
sin, the Upper Peninsula of Michigan (especially on
the Keweenaw Peninsula and the area near White-
fish Point), and in northwestern Wisconsin (Figure
4-12). The highest number (612), largest area
(9,167 ha), and highest areal percentage (9%) of
lakes with ANC ^50 fieq i_~1 were estimated for
Northcentral Wisconsin (2C; Table 4-10). In North-
eastern Minnesota (2A) and Northcentral Wiscon-
Figure 4-11. Classes of ANC (/ueq L"1) in lakes sampled in Region 1 (Northeast), Eastern Lake Survey-Phase I. (Symbols
appearing offshore from Subregion 1E designate lakes sampled on islands.)
51
-------
Table 4-9. Population Estimates of Lakes with ANC =£ 0 jteq
L"1, Eastern Lake Survey- -Phase I
PC
Ncu gc
SUBREGION 1A 0.107 138 190 0.017 2056 3078
1B 0.053 78 141 0.109 2937 5888
1C 0.024 35 63 0.016 1152 2664
1D 0.050 66 107 0.042 1516 3084
1E 0.005 8 21 0.001 95 242
REGION
0.046 326 422 0.018 7756 11567
SUBREGION 2A 0.000 0 (-) 0.000 0 (-)
2B 0.098 102 157 0.024 829 1267
2C 0.031 45 73 0.007 684 1202
2D 0.000 0 (-) 0.000 0 (-)
REGION
0.017 148 209 0.003 1513 2191
SUBREGION 3A 0.000 0 (-) 0.000 0 (-)
3B 0.220 463 615 0.202 13348 18790
pc = estimated proportion of lakes with ANC s 0 jieq L~1.
NC = estimated number of lakes with ANC & 0 n-eq L~1.
NCU = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with ANC £ 0 p,eq L"1.
AC = estimated area of lakes with ANC s 0 (teq L"1.
Acu = 95% upper confidence limit for AC.
(-) = undefined.
Table 4-11. Population Estimates of Lakes with ANC < 200
peg L~1, Eastern Lake Survey -Phase I
PC
9c
SUBREGION 1A 0.705 909 995 0.737 87572 121699
1B 0.387 572 719 0.508 13647 18060
1C 0.676 1002 1100 0.744 53894 72587
1D 0.573 755 882 0.625 22763 28688
1E 0.668 1020 1121 0.632 109546 142497
REGION
1
0.600 4258 4513 0.672 287422 338943
SUBREGION 2A 0.570 830 938 0.233 33277 41006
2B 0.417 438 536 0.172 5842 7292
2C 0.567 839 936 0.145 14171 16904
2D 0.313 1411 1842 0.102 23083 32461
REGION
0.414 3518 3982 0.152 76374 88914
SUBREGION 3A 0.343 88 108 0.546 13258 16436
3B 0.551 1156 1413 0.641 42434 55674
pc = estimated proportion of lakes with ANC == 200 p,eq L"1-
NC = estimated number of lakes with ANC =£ 200 (xeq L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with ANC £ 200 u.eq
L-1.
AC = estimated area of lakes with ANC s 200 |xeq L~1.
ACU = 95% upper confidence limit for A,..
Table 4-10. Population Estimates of Lakes with ANC £50
L"1, Eastern Lake Survey -Phase I
Pc
9c
SUBREGION 1A 0.356 459 541 0.122 14504 21856
1B 0.131 194 281 0.213 5719 8923
1C 0.177 262 334 0.090 6501 9294
1D 0.216 284 368 0.156 5686 8319
1E 0.108 165 220 0.066 11428 21083
REGION
0.192 1364 1536 0.102 43839 56963
SUBREGION 2A 0.042 60 100 0.005 705 1101
2B 0.189 198 265 0.062 2099 2836
2C 0.414 612 705 0.094 9167 11216
20 0.098 441 708 0.021 4766 7943
REGION
0.154 1312 1605 0.033 16737 20609
SUBREGION 3A 0.014 4 5 0.003 79 108
3B 0.353 742 942 0.435 28778 41420
pc = estimated proportion of lakes with ANC =s 50 (xeq L~1.
Nc = estimated number of lakes with ANC =£ 50 (jieq L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with ANC s 50 jteq L"1.
Ac = estimated area of lakes with ANC •& 50 tteq L"1.
ACU = 95% upper confidence limit for AC.
sin (2C), the majority (57%) of lakes was estimated
to have ANC <200 p,eq L"1, although the Upper
Great Lakes Area (2D) also had a large number of
lakes (1,411) in this class (Table 4-11).
4.5.1.4 Southeast
Of all subregions surveyed, Florida (3B) was esti-
mated to have the highest percentage (22%) of
lakes with ANC <0 ixeq L~1 (Table 4-9). Lakes in the
Southeast with ANC ^50 jxeq L 1 were commonly
located in the highland area of the Florida Panhan-
dle, southeast Georgia (Okefenokee Swamp), and
the central peninsula of Florida (Figure 4-13). No
acidic lakes were estimated for the Southern Blue
Ridge (3A) (Table 4-9); the smallest number (4),
least area (79 ha), and lowest numerical and area!
percentages (1% and 0.3%, respectively) of lakes
with ANC =£50 |i,eq L~1 of any subregion occurred
here (Table 4-10). Florida (3B) was also estimated to
have the most lakes with ANC <50 jxeq L~1 (35% or
742 lakes comprising 28,778 ha; Table 4-10). Fifty-
five percent of the lakes in Florida had ANC
s200 u-eq L"1. Similarly, 55 percent of the lake area
in the Southern Blue Ridge and 64 percent of the
lake area in Florida had ANC <200 u,eq L~1 (Table
4-11).
4.5.1.5 Estimates by State for ANC
Population estimates were calculated for each state
where more than 10 lakes were sampled. These
estimates refer only to the portions of states cov-
ered by the ELS-I shown in Section 2.2.1 (Figures
2-1 to 2-4).
No lakes were estimated to have ANC sO (ieq L~1 in
Minnesota, North Carolina, South Carolina or Ver-
mont (Table 4-12). Twenty or fewer lakes were esti-
mated to have ANC =£0 jieq L"1 in Georgia, Maine,
New Hampshire, Pennsylvania and Rhode Island.
The highest numbers of acidic lakes were estimated
for Michigan (107), New York (168) and Florida
(453).
52
-------
Figure 4-12. Classes of ANC (/jeq L~1) in lakes sampled in Region 2 (Upper Midwest), Eastern Lake Survey- -Phase I. (Symbols
appearing offshore from Subregion 2A designate lakes sampled on islands.)
Within the Northeast, the largest estimated number
of lakes with ANC =£0 jxeq L~1 was for New York
(168), followed by Massachusetts (52) and Connect-
icut (47). All other northeastern states were esti-
mated to have 20 or fewer lakes with ANC =£0 jjieq
L"1. In the Upper Midwest, the estimate of acidic
lakes for Michigan (107) exceeded the estimates for
Wisconsin (41) and Minnesota (0). The states in the
Southern Blue Ridge were estimated to have very
few or no acidic lakes, while the highest of all state
estimates of acidic lakes was for Florida (453).
4.5.1.6 Stratified and Shallow Lakes
Of the estimated population of the lakes <2000 ha,
3 percent were weakly stratified and 2 percent were
strongly stratified (Section 2.6.2). In addition, some
lakes were =£3 m in depth and were sampled at
0.5m instead of 1.5 m. Data from these stratified
and shallow lakes were included in population esti-
mates with data from nonstratified lakes sampled
at 1.5 m, because including them had little, if any,
influence on the estimated population characteris-
tics for ANC.
Table 4-13 shows population estimates by subre-
gion for lakes <200 jxeq L"1 ANC based on several
subpopulations of lakes. Listed are nonstratified
lakes sampled at 1.5 m (a), all nonstratified lakes
(b), all stratified lakes (c), lakes sampled at 1.5 m (d),
lakes sampled at 0.5 m (e), and all lakes (f). The
estimates for ANC £200 (xeq L 1 are given as an
example; similar patterns were found for ANC =£0
and £50 (xeq L'1. Some differences were observed
in Subregions 1E, 2A, and 3B. In Maine (1E), the
estimate based on stratified lakes (c) was higher
and the estimate based on shallow lakes (e) was
lower than the estimate for nonstratified lakes sam-
pled at 1.5 m (a). In Northeastern Minnesota (2A),
the estimates based on the same two subpopula-
tions (c and e) were both higher than those for
mixed lakes sampled at 1.5 m (a); but when all lakes
were included (f), the proportion changed only
from 0.524 to 0.570. Shallow lakes (e) in Florida (3B)
S3
-------
Figure 4-13. Classes of ANC (fjeq L*1) in lakes sampled in Subregions 3A (Southern Blue Ridge) and 3B (Florida), Eastern Lake
Survey-Phase I.
/
had higher ANC than lakes sampled at 1.5 m (d),
producing a decrease in the proportion of lakes
with ANC <200 jxeq L~1 from 0.608 for lakes sam-
pled at 1.5 m (d) to 0.551 for all lakes (f).
In general, estimates of the proportion of lakes with
ANC s200 |xeq L"1 based on the subpopulations
(a - e) were very close to the estimates for all lakes
(f). Adding shallow (e) and stratified (c) lakes to the
unstratified lakes sampled at 1.5 m (a) had little
effect on the population estimates of the proportion
of lakes for most subregions. Three subregions (1B,
2B, and 3B) show differences >0.05 in the propor-
tions of lakes with ANC =s200 p,eq L~1 between non-
stratified (a) and all lakes (f). Subregions 2B and 3B
contain a large population of shallow lakes (Table
54
-------
Table 4-12. Estimates of Numbers of Lakes with ANC <0,
<50 and <200 /jeq L"1 by State8, Eastern Lake
Survey-Phase I
Estimated
Number of Number of
Lakes Lakes
State (N) Sampled
CT
FL
GA
MA
ME
Ml
MN
NC
NH
NY
PA
Rl
SC
VT
Wl
346
2088
155
926
1966
2073
3026
55
639
2041
616
113
40
258
3402
24
138
54
97
225
160
174
30
69
191
106
15
12
29
253
ANC (n,eq L~l)
<0 (UCL)"
47 (100)
453 (605)
10 (10)
52 (83)
8(21)
107 (162)
0(-)c
0(-)c
17 (35)
168 (237)
20 (30)
13 (27)
0(-)0
0(-)c
41 (68)
<50 (UCL)0
47 (100)
732 (932)
10 (10)
239(311)
200 (261)
368 (522)
143 (282)
4(5)
171 (229)
577 (693)
79 (95)
33 (54)
0(-)°
19 (36)
801 (1008)
<200 (UCL)"
145 (218)
1146(1403)
49(63)
578 (685)
1337 (1450)
704 (910)
1124(1368)
35(48)
537 (609)
1200 (1349)
284 (363)
86(121)
10(17)
90(121)
1690 (2027)
'Includes only states, In which more than ten lakes were sampled.
"Upper confidence limit, Ncu, shown in parentheses.
c(-j = undefined.
4-7) that strongly influence these proportions.
Therefore, eliminating lakes not originally desig-
nated in the protocol for sampling (i.e., shallow and
stratified lakes) would substantially reduce the esti-
mated number of lakes with ANC <200 (jieq L~1 in
Subregion 1B from 572 (f) to 332 (a), in Subregion
2B from 438 (f) to 245 (a), and in Subregion 3B from
1156 (f) to 1042 (a). To avoid loss of information on
these subpopulations, all population estimates re-
ferred to in this report include lakes regardless of
thermal stratification status or sample depth.
Table 4-13. Population Estimates of the Proportion of Lakes
with ANC £200 peq L~1 for Six Subpopula-
tions,* Eastern Lake Survey -Phase I
ANC s200 M-eq L~1
SUBREGION
1A
1B
1C
1D
1E
0.703
0.328
0.675
0.585
0.703
0.706
0.349
0.671
0.576
0.662
0.696
(0.000)**
0.741
(0.000)**
(0.826)
0.702
0.381
0.681
0.581
0.709
0.738
0.404
0.655
0.554
0.496
0.705
0.387
0.676
0.573
0.668
REGION
1 0.604 0.589 0.779
0.617 0.537
0.600
SUBREGION 2A 0.524 0.561 (0.767) 0.536 0.744 0.570
2B 0.471 0.412 0.449 0.466 0.335 0.417
2C 0.577 0.562 (0.000)** 0.582 (0.000)** 0.567
2D 0.302 0.324 (0.000)** 0.291 0.442 0.313
REGION
0.414 0.420 0.338
0.407 0.451
0.414
SUBREGION 3A 0.371 0.327 (0.000)** 0.388 (0.000)** 0.343
3B 0.608 0.551 0.608 0.296 0.551
*a = nonstratified lakes sampled at 1.5 m
b = all nonstratified lakes
c = all stratified lakes
d = lakes sampled at 1.5 m
e = lakes sampled at 0.5 m (shallow)
f = all lakes
"Proportions in parentheses are based on sample sizes of <10;
in Subregion 3B, no stratified lakes were sampled.
4.5.2 pH
4.5.2.1 Reference Values
The two reference values chosen for pH were 5.0
and 6.0. A pH of =s5.0 was defined as low pH. The
tables for pH <6.0 are cumulative; i.e., they include
lakes <5.0.
4.5.2.2 Northeast
In the Northeast, lakes with low pH were most com-
mon in the Adirondacks, Poconos, Cape Cod, and
southwestern New Hampshire (Figure 4-14). The
highest number (128) and percentage (10%) of low
pH lakes were estimated for the Adirondacks (1 A);
however, the largest area (2,295 ha) and areal per-
centage estimates (6%) of low pH lakes were for
Southern New England (1D; Table 4-14). Maine (1E)
had few low pH lakes (8) and had the fewest (74)
lakes with pH <6.0 (Table 4-15). The highest num-
ber (343) and percentage (27%) and the largest area
(12,375 ha) of lakes with pH <6.0 were also found in
the Adirondacks (1 A), but the highest areal percent-
age (14%) of lakes with pH s6.0 occurred in the
Poconos/Catskills(1B).
4.5.2.3 Upper Midwest
Lakes in the Upper Midwest with low pH were most
common in Northcentral Wisconsin (2C) and the
Upper Peninsula of Michigan (2B) near Whitefish
Point (Figure 4-15). The highest number (99), great-
est area (812 ha), and corresponding percentages
(9% and 2%) of low pH lakes were estimated to
occur in the Upper Peninsula of Michigan (2B). No
low pH lakes were sampled in Subregion 2A or 2D
(Table 4-14). The highest number (411), largest area
(6,276 ha), and highest percentages for both num-
ber (28%) and area (6%) of lakes with pH <6.0 were
estimated for Northcentral Wisconsin (2C). North-
eastern Minnesota (2A) had the fewest number
(20), least area (255 ha), and lowest corresponding
percentages (1% and 0.2%) of lakes with pH <6.0
(Table 4-15).
4.5.2.4 Southeast
In the Southeast, lakes with low pH were located in
the highland region of the Florida Panhandle,
southern Georgia, and the central ridge of the Flor-
ida Peninsula (Figure 4-16). No low pH lakes were
sampled in the Southern Blue Ridge (3A) and only
one lake with pH <6.0 (Tables 4-8 and 4-9). Florida
(3B) was estimated to have the most low pH lakes
(12% or 259 lakes comprising 7,936 ha) and the
most lakes with pH <6.0 of any ELS-I subregion
(33% or 687 lakes comprising 21,635 ha).
4.5.2.5 Estimates by State for pH
Population estimates for lakes with pH <5.0 and
<6.0 were calculated for each state. As with ANC,
55
-------
Figure 4-14. Classes of pH in lakes sampled in Region 1 (Northeast), Eastern Lake Survey-Phase I. (Symbols appearing offshore
from Subregion 1E designate lakes sampled on Islands.)
Symbol
•
+
o
Value
<5
5-6
>6
Table 4-14.
Population Estimates of Lakes with pH < 5.0,
Eastern Lake Survey-Phase I
9c
Acu
SUBREGION 1A 0.100 128 179 0.017
1B 0.008 12 20 0.019
1C 0.017 25 46 0.003
1D 0.050
1E 0.005
66 114 0.063
8 21 0.001
1962 2977
513 847
194 372
2295 4555
95 242
REGION
1
0.034 240 314 0.012 5059 7570
SUBREGION 2A 0.000 0 (-) 0.000 0 (-)
2B 0.094 99 153 0.024 812 1251
2C 0.021 31 55 0.003 252 468
2D 0.000 0 (-) 0.000 0 (-)
REGION
0.015 130 189 0.002 1064 1553
SUBREGION 3A 0.000 0 (-) 0.000 0 (-)
3B 0.124 259 385 0.120 7936 12666
pc = estimated proportion of lakes with pH -s, 5.0.
Nc = estimated number of lakes with pH s 5.0.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with pH < 5.0.
Ac = estimated area of lakes with pH £ 5.0.
ACU = 95% upper confidence limit for A,..
(-) = undefined.
these estimates refer only to the portions of the
states covered by the ELS-I shown in Section 2.2.1
(Figures 2-1 to 2-4). No lakes with pH <5.0 were
sampled in Minnesota, North Carolina, Rhode Is-
land, South Carolina and Vermont (Table 4-16).
More than 100 lakes in this category were esti-
mated for three states: Florida (249), Michigan (103)
and New York (128). States with the highest num-
bers of lakes estimated to have pH ^6.0 were: Flor-
ida (677), Massachusetts (180), Michigan (330),
Minnesota (103), New Hampshire (126), New York
(384) and Wisconsin (386).
4.6 Regional and Subregional Population
Estimates for Other Primary Variables
4.6.1 Sulfate
The data from the ELS-I were examined and two
reference values for SO/2 were arbitrarily selected
for comparison, >50 and >150 /ueq L~1. These ref-
erence values were not selected to represent or imply
a background or critical level of SO/2, but are
discussed to illustrate differences among areas.
56
-------
Figure 4-15. Classes of pH in lakes sampled in Region 2 (Upper Midwest), Eastern Lake Survey -Phase I. (Symbols appearing
offshore from Subregion 2A designate lakes sampled on islands.)
Symbol
•
+
0
Value
<5
5-6
>6
Table 4-15.
Population Estimates of Lakes with pH < 6.0,
Eastern Lake Survey-Phase I
9c
AQ A50 and
>150 //eq L'1. The 1 -F(x) and 1 -G(x) curves for SCV2
are found in Volume II.
Within the Northeast, nearly all lakes were estimated
to have SO/2 concentrations >50 //eq L"1 (Table 4-
17). More than 99 percent of all lakes in Subregions
1A-1D had SO4"2 values >50//eq L"1. In Maine (1E),
88.5 percent of the lakes were estimated to have
S04~2>50//eqL.-1.
The Poconos/Catskills (1B) were estimated to have
the largest number of lakes (829, 56%) with S04~2
concentrations >150 //eq L"1 (Table 4-18). Southern
New England (1D) also had a large number of lakes
(604 or 46%) with S04"2 concentrations >150//eq L"1.
Only 16 lakes (1 %) in Maine (1E), were estimated to
have SO/2 concentrations >150 //eq L"1. The
Adirondacks (1A) had the largest lake area (21,987
-1
ha) with SO4 concentrations >150 //eq L
In the Upper Midwest, the largest number of lakes
(495, 11%) with SO4~2 >150 //eq L"1 was located in
57
-------
Figure 4-16.
Classes of pH in lakes sampled in Subregions 3A (Southern Blue Ridge) and 3B (Florida). Eastern Lake Survey -
Phase I.
Symbol
•
+
o
Value
<5
5-6
>6
the Upper Great Lakes Area (2D; Table 4-18). North-
central Wisconsin (2C) had the lowest estimated
number (16, 1%) of lakes with SO4"2 ^150 /ueq L"1.
The largest area of lakes having S04"2 ^150 fjeq L"1
was found in Subregion 2B, the Upper Peninsula of
Michigan (10,889 ha, 32%). Northcentral Wisconsin
had the smallest lake area (445 ha) with S04~2 >150
yueq L"1.
The Southern Blue Ridge (3A) contained very few
lakes (22, 8%) and a small lake area (1,720 ha, 7%)
58
with S04"2 >150 //eq L"1 (Table 4-18). Estimates for
Florida (3B) gave the largest number of lakes (846,
40%) and the largest lake area (30,443 ha, 46%) with
S04"2 >150 yueq L"1 of any subregion.
4.6,2 Calcium
Three reference values originally were selected for
calcium based on examination of the ELS-I data.
The F(x) and G(x) distributions for calcium, with the
statistics for the three reference values, are in-
-------
Table 4-16. Estimates of Numbers of Lakes with pH <5.0
and <6.0 by State*, Eastern Lake Survey-Phase
Estimated
Number of
Lakes
State (N)
CT
FL
GA
MA
ME
Ml
MN
NC
NH
NY
PA
Rl
SC
VT
Wl
346
2088
155
926
1966
2073
3026
55
639
2041
616
113
40
258
3402
Number of
Lakes
Sampled
24
138
54
97
225
160
174
30
69
191
106
15
12
29
253
PH
<5.0 (UCL)"
19 (50)
249 (375)
10 (10)
54 (93)
8(21)
103 (158)
0(-)e
0(-)°
17 (35)
128 (179)
13 (20)
0(-)c
0(-)°
0(-)c
27(49)
<6.0 (UCL)b
47 (100)
677 (868)
10(10)
180 (243)
90 (132)
330 (482)
103 (238)
1(2)
126 (177)
384 (473)
58 (73)
20 (36)
0(-)°
11 (27)
386 (465)
"Includes only states in which more than ten lakes were sampled.
"Upper confidence limit, Neu, shown in parentheses.
°(-) = undefined.
Table 4-17.
Population Estimates of Lakes with Sulfate s50
ixeq L~1, Eastern Lake Survey-Phase I
PC
Nr.
9c
SUBREGION 1A 0.994 1282 1361 0.999 118664 155162
1B 0.998 1476 1628 0.998 26818 31611
1C 0.995 1475 1570 0.999 72321 92086
1D 1.000 1318 1472 1.000 36403 43610
1E 0.885 1350 1463 0.950 164696 206822
REGION
1
0.973 6901 7175 0.979 418903 478671
SUBREGION 2A 0.737 1073 1204 0.878 125600 186914
2B 0.791 831 954 0.943 32071 49992
2C 0.695 10^8 1138 0.831 81078 115637
2D 0.501 2261 2797 0.357 81029 139434
REGION 2 0.611 5193 5768 0.638 319791 412988
SUBREGION 3A 0.231
3B 0.681
59 82 0.300
1428 1720 0.783
7281
51787
13390
67699
pc = estimated proportion of lakes with sulfate a50 ixeq L~1.
NC = estimated number of lakes with sulfate £50 jieq L"1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with sulfate a50 j.i,eq
L-1.
AC = estimated area of lakes with sulfate a50 u.eq L"1.
ACU = 95% upper confidence limit for Ac.
eluded in Volume II. In this volume only the refer-
ence value of Ca+2 =£50 M-eq L~1 is discussed. Table
4-19 summarizes the population estimates for Ca+2
derived using this reference value.
In the Northeast, Southern New England (1D) was
estimated to contain the largest number of lakes
(133,10%) and the largest lake area (2,164 ha, 6%)
having Ca+2 concentrations ^50 u-eq L~1 (Table
4-19). The Poconos/Catskills(1 B) had very few lakes
with Ca+2 concentrations at or below the reference
value of ^50 u,eq L~1.
Table 4-18.
Population Estimates of Lakes with Sulfate &150
(teq L-1, Eastern Lake Survey -Phase I
PC
Nc
AC Acu
SUBREGION
REGION
SUBREGION
REGION
SUBREGION
1A
1B
1C
1D
1E
1
2A
28
2C
2D
2
3A
3B
0.128
0.561
0.157
0.458
0.011
0.260
0.019
0.066
0.010
0.110
0.071
0.085
0.403
165
829
232
604
16
1846
28
69
16
495
608
22
846
216
1003
299
737
34
2082
60
123
40
813
932
36
1088
0.185
0.533
0.160
0.518
0.006
0.158
0.012
0.320
0.005
0.024
0.037
0.071
0.460
21987
14330
11577
18871
1022
67788
1661
10889
445
5358
18353
1720
30443
35973
18076
17882
25309
2488
84904
3555
26780
1154
9234
34835
2737
44423
pc = estimated proportion of lakes with sulfate >150 (j,eq L"1.
Nc = estimated number of lakes with sulfate a 150 u,eq L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with sulfate a 150 u,eq
L-1.
Ac = estimated area of lakes with sulfate >150 (ieq L"1.
ACU = 95% upper confidence limit for A^
Table 4-19.
Population Estimates of Lakes with Calcium <50
(ieq L~1, Eastern Lake Survey -Phase I
PC
Nc
9c
Acu
SUBREGION 1A 0.083 108 154 0.014 1618 2579
1B 0.004 6 11 0.003 93 154
1C 0.042 63 100 0.021 1534 3225
ID 0.101 133 183 0.059 2164 3763
1E 0.032 49 80 0.003 561 956
REGION
1 0.051 359 442 0.014 5970 8520
SUBREGION 2A 0.019 28 60 0.002 234 495
2B 0.160 168 235 0.044 1496 2177
2C 0.219 324 400 0.047 4608 6211
2D 0.057 256 450 0.016 3564 6451
REGION
0.091 776 997 0.020 9902 13284
SUBREGION 3A 0.120 31 42 0.203 4927 6858
3B 0.192 402 551 0.186 12289 17763
PC = estimated proportion of lakes with calcium £50 fieq L~1.
Nc = estimated number of lakes with calcium ==50 M-eq L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of lake .area with calcium £50 jieq
L-1.
AC = estimated area of lakes with calcium =£50 jieq L"1.
ACU = 95% upper confidence limit for AC-
For estimates within the Upper Midwest, Northcen-
tral Wisconsin (2C) had the largest number of lakes
(324,22%) and largest lake area (4,608 ha, 5%) with
Ca+2 values <50 u,eq L~1 (Table 4-19). The Upper
Peninsula of Michigan (2B) also had a high percent-
age (16%) of lakes with Ca+2 concentrations <50
(jueq L~1. The Upper Great Lakes Area (2D) had 256
lakes (6%) and a lake area of 3,564 ha (1.6%) with
Ca+2 values ^50 u,eq L~1. Few lakes with low Ca+2
-------
concentrations were located in Northeastern Min-
nesota (2A).
In the Southern Blue Ridge (3A), 12 percent of the
lakes and 20 percent of the lake area were esti-
mated to have Ca+2 <50 u,eq |_~1 (Table 4-19). Flor-
ida (3B) had the most lakes (402,19%) of any subre-
gion with Ca+2 values <50 jieq L~1.
4.6.3 Extractable Aluminum
Monomeric aluminum (Al) can be toxic to fish at
concentrations as low as 100-200 u-9 L~1 unless
high concentrations of chelators are also present
(Driscoll et al. 1980; Baker and Schofield 1982; Baker
1984). Darkwater and clearwater lakes are defined
here as having true color >30 PCU and <30 PCU,
respectively. For this report, it was assumed that
darkwater lakes have sufficient organic material to
complex monomeric forms of aluminum. Therefore,
population estimates based on reference values for
extractable Al are presented for clearwater lakes
only. Reference values 2=50,5=100, and >150 u,g L~1
were selected to bracket the lowest toxic concentra-
tions cited above. Cumulative frequency and areal
distributions [1 -F(x) and 1 -G(x)] (Section 4.3.2.1) for
extractable aluminum are provided in Volume II.
The largest estimated number of clearwater lakes
(120, 14%) having extractable Al concentrations
s50 u,g L"1 (Table 4-20) occurred in the Adiron-
dacks (1 A). No clearwater lakes with concentrations
exceeding 50 u,g L~1 were estimated for Maine (1 E),
Northeastern Minnesota (2A), the Upper Great
Lakes Area (20), or the Southern Blue Ridge (3A).
Other subregions contained a small percentage of
clearwater lakes, ranging between 1 and 7% with
extractable Al in excess of 50 (xg L~1.
The largest number of lakes having extractable Al
concentrations >100 u,g L~1 also occurred in the
Adirondacks (100, 12%). Four additional subre-
gions, Poconos/Catskills (1 B), Southern New Eng-
land (1D), the Upper Peninsula of Michigan (2B),
and Florida (3B), were estimated to have small per-
centages of lakes with extractable Al >100 u,g L~1.
In Subregion 1A, 82 (10%) of the lakes were esti-
mated to have concentrations of extractable Al
&150 jig I"1 (Table 4-21). Other subregions having
clearwater lakes with extractable Al concentrations
>150 /ug L'1 (Table 4-22) occurred in the Poconos/
Catskills (1 B, 3) Southern New England (1 D, 7), the
Upper Peninsula of Michigan (2B, 2), and Florida
(3B, 14). No clearwater lakes in other subregions
had extractable Al concentrations >150 u,g L"1.
4.6.4 Dissolved Organic Carbon
Reference values of 2 mg L~1 and 6 mg L~1
were
used to characterize lakes with respect to DOC. In
lakes with DOC concentrations &6 mg L~1, organic
anions are likely to be important in contributing to
Table 4-20.
Population Estimates of Clearwater Lakes with
Extractable Aluminum >50 |ig L~\ Eastern Lake
Survey -Phase I
PC
Nc
9c
AC
SUBREGION 1A 0.144 120 169 0.025
1B 0.060 66 128 0.127
1C 0.018
1D 0.040
1E 0.000
19
33
0
39 0.016
67 0.044
(-) 0.000
REGION
1
0.052 237 325 0.025
2072 3124
2787 5738
1021 2525
1079 2614
0 (-)
6959 10758
SUBREGION 2A
2B
2C
2D
REGION 2
SUBREGION 3A
3B
0.000
0.021
0.009
0.000
0.005
0.000
0.074
0
11
8
0
19
0
69
(-)
17
21
(-)
33
(-)
114
0.000
0.015
0.001
0.000
0.001
0.000
0.127
0
196
44
0
240
0
4600
(-)
318
112
(-)
380
(-)
8750
pc = estimated proportion of clearwater lakes with extractable
aluminum >50 jig L~1.
Nc = estimated number of clearwater lakes with extractable
aluminum a50 jig L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of clearwater lake area with ex-
tractable aluminum £50 |ig L"1.
AC = estimated area of clearwater lakes with extractable alu-
minum >50 jig L~1.
Acu = 95% upper confidence limit for AC.
(-) = undefined.
60
Table 4-21. Population Estimates of Clearwater Lakes with
Extractable Aluminum >100 fig L'1, Eastern
Lake Survey -Phase I
Me "c "cu «c «c «cu
SUBREGION 1A
1B
1C
1D
1E
REGION 1
SUBREGION 2A
2B
2C
2D
0.121
0.029
0.000
0.008
0.000
0.030
0.000
0.004
0.000
0.000
100
32
0
7
0
139
0
2
0
0
146
76
(-)
16
(-)
203
(-)
4
(-)
(-)
0.022
0.098
0.000
0.002
0.000
0.014
0.000
0.007
0.000
0.000
1784
2149
0
40
0
3973
0
94
0
0
2789
4966
(-)
101
(-)
6965
(-)
199
(-)
(-)
REGION
0.000
SUBREGION 3A 0.000
3B 0.044
0
40
4 0.000
(-) 0.000
76 0.064
94
0
2324
199
5656
pc = estimated proportion of clearwater lakes with extractable
aluminum >100 |xg L"1.
Nc = estimated number of clearwater lakes with extractable
aluminum a 100 |ig L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of clearwater lake area with ex-
tractable aluminum >100 |xg L~1.
AC - estimated area of clearwater lakes with extractable alu-
minum &100 |ig L"1.
ACU = 95% upper confidence limit for A,-.
(-) = undefined.
-------
Table 4-22.
Population Estimates of Clearwater Lakes with
Extractable Aluminum >150 |ig L~1, Eastern
Lake Survey-Phase I
PC
NC
NC
gc
SUBREGION 1A 0.099 82 124 0.017 1439 2377
1B 0.003 3 8 0.005 101 239
1C 0.000 0 (-) 0.000 0 (-)
1D 0.008 7 16 0.002 40 101
1E 0.000 0 (-) 0.000 0 (-)
REGION
1
0.020 92 135 0.006 1580 2531
SUBREGION 2A
2B
2C
2D
REGION 2
SUBREGION 3A
3B
0.000
0.004
0.000
0.000
0.000
0.000
0.015
0
2
0
0
2
0
14
(-)
4
(-)
(-)
4
(-)
35
0.000
0.007
0.000
0.000
0.000
0.000
0.003
0
94
0
0
94
0
103
(-)
199
(-)
(-)
199
(-)
229
pc = estimated proportion of clearwater lakes with extractable
aluminum >150 |j,g L"1.
Nc = estimated number of clearwater lakes with extractable
aluminum a150 p,g L~1.
Ncu = 95% upper confidence limit for Nc.
gc = estimated proportion of clearwater lake area with ex-
tractable aluminum >150 n,g L~1.
Ac = estimated area of clearwater lakes with extractable alu-
minum >150 jig L~1.
Acu = 95% upper confidence limit for AC.
(-) = undefined.
the sum of anions (Oliver et al. 1983). The ion bal-
ance in lakes having DOC concentrations of 2 to
6mg L~1 can be dominated either by organic
anions or bicarbonate or carbonate ions, depend-
ing on pH. Population estimates using these refer-
ence values (^2 mg L"1 and ^6 mg L~1) are pre-
sented in Tables 4-23 and 4-24. Cumulative
frequency and area! distributions [F(x) and G(x);
1-F(x) and 1-G(x)] are presented in Volume II.
Southern New England (1D) contained the most
lakes (238,18%) in the Northeast with DOC concen-
trations <2 mg L~1 (Table 4-23). In other areas of
the Northeast, only 4 to 7 percent of the lakes were
estimated to have DOC <2 mg L~1.
The Upper Midwest generally had fewer lakes with
DOC concentrations <2 mg L~1 than the Northeast.
Only three lakes (0.2%) in Northeastern Minnesota
(2A) had DOC concentrations ^2 mg L~1. Two lakes
in Subregion 2D were estimated to have DOC ^2
mg L~1; only 4 percent of the lakes in Subregions
2B and 2C were estimated to have DOC <2 mg L~1.
Florida (3B) contained the highest number of lakes
(242) with DOC concentrations =£2 mg L~1. The
highest percentage of lakes with DOC ^2 mg L~1
occurred in the Southern Blue Ridge (3A, 54%). The
Southern Blue Ridge contained the largest area
(21,255 ha, 88%) of lakes with DOC concentrations
<2 mg L~1.
Table 4-23. Population Estimates of Lakes with DOC s2 mg
L~1, Eastern Lake Survey -Phase I
PC
Nc
NBI
9c
SUBREGION 1A 0.074 95 138 0.063 7518 14564
1B 0.055 81 143 0.151 4066 7718
1C 0.062 92 136 0.065 4682 8416
1D 0.180 238 318 0.158 5740 8469
1E 0.044 67 103 0.060 10321 22261
REGION 1 0.081 572 696 0.076 32327 47392
SUBREGION 2A 0.002 3 7 0.000 64 150
2B 0.044 46 85 0.023 777 1466
2C 0.041 61 98 0.008 777 1304
2D 0.000 2 3 0.000 37 73
REGION
0.013 111 165 0.003 1654 2528
SUBREGION 3A 0.536 138 164 0.876 21255 29067
3B 0.115 242 340 0.141 9327 14289
pc = estimated proportion of lakes with DOC ==2 mg L~1.
NC = estimated number of lakes with DOC <2 mg L~1.
NCU = 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with DOC ==2 mg L~1.
AC = estimated area of lakes with DOC <2 mg L~1.
ACU = 95°/° upper confidence limit for AC-
Table 4-24. Population Estimates of Lakes with DOC >6 mg
L"1 Eastern Lake Survey' -Phase I
Pc NC Ncu gc AC ACU
SUBREGION 1A 0.149 192 249 0.038 4493 6347
1B 0.186 274 398 0.121 3247 4857
1C 0.275 408 494 0.143 10353 14556
1D 0.270 356 462 0.224 8167 12070
1E 0.421 643 743 0.452 78303 109175
REGION 1 0.264 1873 2090 0.244 104563 136060
SUBREGION 2A 0.760 1107 1238 0.509 72732 101387
2B 0.567 595 713 0.840 28590 46581
2C 0.342 507 609 0.133 12988 20160
2D 0.696 3141 3692 0.491 111306 162583
REGION
0.629 5351 5938 0.450 225616 287467
SUBREGION 3A 0.061 16 28 0.016 393 680
3B 0.689 1445 1766 0.620 41004 55386
pc = estimated proportion of lakes with DOC 2:6 mg L"1.
Nc = estimated number of lakes with DOC &6 mg L~1.
Ncu - 95% upper confidence limit for Nc.
gc = estimated proportion of lake area with DOC a6 mg L~1.
A,. = estimated area of lakes with DOC a6 mg L~1.
Acu = 95% upper confidence limit for AC-
Maine (1E) had the highest percentage of lakes
(42%) with DOC concentrations >6 mg L"1 in the
Northeast, followed by Subregions 1C and 1D with
27.5 percent and 27.0 percent, respectively (Table
4-24). The Adirondacks (1 A) had the lowest percent-
age of high DOC lakes (15%) in the Northeast.
A large number of lakes with DOC concentrations
s=6 mg L~1 occurred in the Upper Midwest. The
estimated percentage of the total number of such
61
-------
lakes in subregions of the Upper Midwest ranged
from 34 percent in Northcentral Wisconsin (2C) to
76 percent in Northeastern Minnesota (2A). The
Southern Blue Ridge (3A) contained the smallest
percentage (6%) of lakes having high DOC concen-
trations (>6 mg L~1) of any ELS-I subregion. Florida
(3B) contained the second highest number of lakes
with DOC concentrations >6 mg L"1 (1445, 69%).
4.7 Statistics for Population
Distributions, Primary Variables
Table 4-25 gives the population estimates of the
median values for the primary variables: pH, ANC,
DOC, extractable Al, S04~2 and Ca+2. The first and
fourth quintile values (Q, and Q4) of the distribution
are also given. These quintiles represent the 20th
and 80th percentiles of the distribution, respec-
tively.
Median, Q1 and Q4 values are used to compare the
distribution, F(x), among subregions and regions.
In comparing any two distributions, for example, it
is unlikely that the values of these three statistics
would be the same if the two distributions were
Table 4-25. Primary Variables: First Quintiles (Q,), Medians
(M), and Fourth Quintiles (04), Eastern Lake Sur-
vey-Phase!
SUBREGION
REGION
1A
IB
1C
10
IE
1
SUBREGION 2A
REGION
2B
2C
20
2
SUBREGION 3A
3B
Ql
5.48
6.64
6.30
6.21
6.57
6.33
6.57
6.07
5.80
6.63
6.40
6.73
5.38
pH
M
6.71
7.02
6.77
6.81
6.91
6.87
6.94
7.10
6.68
7.39
7.09
6.98
6.56
ANC
-------
respectively, the other subregions had Q4 values
ranging from 9.2 to 13.5 jxg L~1. Similarly, esti-
mated Q4 values of ammonium in Subregions 1B
and 1D were more than twice those in other areas.
Concentrations of ammonium, nitrate and phos-
phorus in Maine (1E) were all low (medians = 1.2
ixeq L~1,0.2 u,eq L~1 and 5.7 u,eq L"1, (respectively).
Total phosphorus was also low in the Adirondacks
(median = 4.7 u-g L"1; Q4 = 9.2 u,g L"1).
Estimated median values of nutrients were some-
what higher in the Upper Midwest than in the
Northeast. Within the Upper Midwest, concentra-
tions were generally highest in the Upper Great
Lakes Area (2D), where the total phosphorus me-
dian was 18.9 u,g L~1. Total phosphorus concentra-
tions were similar among lakes in Subregions 2A-2C
(median values = 12.6 - 13.7 fjg L~1).
Nitrate concentrations in the Southern Blue Ridge
(3A) were higher at the median (3.1 u,eq Lr^than in
any other area. Median ammonium values were as
high in 3A (2.2 jxeq L"1) as in 1B and 2C, and the
estimated Q4 exceeded that in all other subregions
(11.0 jjieq L~'). Total phosphorus was generally low
in 3A.
Estimated median and Q4 nitrate values in Florida
(3B) were less than in the Southern Blue Ridge (3A),
but greater than in the Northeast or Upper Mid-
west. Ammonium concentrations in Florida were
within the range of median and Q4 values estimated
for the Northeast and Upper Midwest, as were con-
centrations of total phosphorus. In Florida, total
phosphorus (median = 12.4 u,g L~1) was greater at
all quintiles than in the Southern Blue Ridge, and
exceeded the regional estimates of median and Q4
values in the Northeast.
4.8.2 True Color, Turbidity and Secchi Disk
Transparency
The population estimates for true color were simi-
lar to those for DOC (Section 4.6.4). With the excep-
tion of Subregion 2C, estimated median values of
true color for the Upper Midwest and Southeast
were greater than those calculated for the North-
east (Table 4-27). True color values in the Northeast
at all levels were strikingly similar, with median
values from 22 to 25 PCU. The highest median color
among subregions in the Midwest (44 PCU) was
observed in Northeastern Minnesota, which had a
Q4 of 89 PCU. The third highest median (37 PCU)
and the highest Q4 (93 PCU) were found in Florida.
Turbidity was generally low and uniform among
most subregions. The highest turbidity was found
in the Southern Blue Ridge (3A), where 90 percent
of the lakes are estimated to be reservoirs. Turbid-
ity was also slightly higher than other subregions in
the Poconos/Catskills (1B), which also contain
Table 4-27. Secondary Variables (True Color, Turbidity and
Secchi Disk Transparency): First Quintiles (Q-i),
Medians (M), and Fourth Quintiles (04), Eastern
Lake Survey-Phase I
Secchi Disk
Transparency
(m)
True Color
(PCU)
Turbidity
(NTU)
Q, M Q4 Q, M Q4 Q, M Q4
SUBREGION
REGION
SUBREGION
REGION
1A
1B
1C
1D
1E
1
2A
2B
2C
2D
2
10
12
14
11
10
13
24
16
13
15
15
22
22
24
24
25
24
44
31
24
39
35
41
33
41
52
46
44
89
74
50
74
76
0.3
0.8
0.3
0.4
0.4
0.4
0.6
0.6
0.4
0.5
0.5
0.5
1.4
0.7
0.9
0.6
0.7
1.1
0.9
0.8
1.0
0.9
0.9
2.8
1.2
2.9
1.0
1.6
1.8
1.6
1.6
2.0
1.9
1.8
1.2
1.2
0.9
1.2
1.2
1.0
0.9
1.3
1.0
1.0
2
1
2
1
2
2
1
1
2
1
1
.8
.8
.6
.6
.6
.3
.6
.5
.3
.9
.9
4.9
2.6
4.4
3.0
4.7
4.0
2.8
2.9
3.4
3.3
3.3
SUBREGION 3A 21 36 59 1.9 3.9 9.1 0.9 1.8 2.7
3B 14 37 93 0.5 0.9 2.2 1.0 1.9 3.0
many reservoirs (Section 4.4.2.8). The Q4 value cal-
culated for Southern New England (1D) was also
greater than in other subregions of the Northeast.
Secchi disk transparency represents the average of
disappearance and reappearance depth of the disk.
Where Secchi disk transparency was equal to the
site depth (i.e., where the disk was still visible on
the lake bottom), no mean value was calculated.
The population estimates for Secchi disk transpar-
ency exclude the measurements made on these
lakes.
Secchi disk transparency at Q-, was 0.9 - 1.3 m in all
subregions, except in the Adirondacks (1A), which
had a Q1 value of 1.8 m. The Adirondacks also had
the highest median (2.8 m) and Q4 (4.9 m) values of
the subregions. Within the Northeast, Secchi disk
transparency was also high in 1C (4.4 m at Q4) and
1E (4.7 m at Q4), and substantially lower (2.6 m and
3.0 m at Cu in 1 B and 1D, respectively.
In the Upper Midwest, estimated Q1f median and Q4
values of Secchi disk transparency were greatest in
Northcentral Wisconsin (2C), which contains many
clearwater seepage systems. The highly colored
lakes of Northeastern Minnesota (and also some
areas of 2B) had the lowest Secchi disk transpar-
ency in the Upper Midwest. Transparencies in the
Southeast were similar to lakes in Subregion 2A.
4.8.3 Sodium, Potassium and Magnesium
Within the Northeast, estimated Q1f median and Q4
sodium concentrations were highest in Southern
New England (1D), and intermediate in 1B and 1C
(Table 4-28). Slightly lower d, and median concen-
63
-------
Table 4-28.
Secondary Variables (Sodium, Potassium and
Magnesium): First Quintiles (Q,), Medians (M),
and Fourth Quintiles (Q4), Eastern Lake
Survey-Phase I
Na+ (M.eq L-1) K+
L-1)
L~i)
1
M 04
,
M
,
M
SUBREGION 1A 23.3 37.5 78.4 5.8 8.5
1B 49.3 103.4 267.0 9.7 17.1
1C 44.1 70.1 253.4 7.5 11.1
1D 150.5 342.3 773.0 16.6 24.7
IE 38.0 62.5 112.0 6.5 9.1
11.6 25.2 53.9 101.3
29.8 61.6 101.1 205.6
22.1 32.0 49.4 111.0
49.1 64.3 116.7 227.0
13.4 35.7 55.9 93.6
REGION
1 38.6 82.8 285.2 7.4 12.2 24.9 37.0 69.6 150.7
SUBREGION 2A
2B
2C
2D
REGION 2
33.6 40.1 52.1 6.5 9.2
13.5 28.9 46.6 8.6 13.4
9.1 30.8 65.4 10.8 14.8
23.3 66.5 118.3 15.2 21.1
13.8 71.2 100.4 165.5
20.0 39.2 148.4 419.9
19.0 29.6 58.0 238.5
30.6 65.5 340.4 700.7
20.3 41.2 95.3 10.2 16.3 25.4 54.3 152.4 445.2
SUBREGION 3A 59.9 103.1 156.1 21.0 39.4 57.6 44.3 76.1 117.0
3B 107.4 197.9 340.1 4.6 21.6 139.6 55.8 159.9 494.5
trations and a much lower 04 concentration were
estimated for Maine (1E). The Adirondacks (1 A) had
the lowest quintile concentrations for sodium in the
Northeast. Estimates of concentrations of sodium
for the Upper Midwest were less than for most
areas of the Northeast and for Subregions 3A and
3B. Subregion 2C had the lowest calculated Q1 con-
centration and 2D had the highest median and Q4
concentrations in the Upper Midwest. Little in-
terquintile difference (Q4-Q.,) for sodium was
found in Northeastern Minnesota (2A). Estimated
median concentrations for sodium in Florida (3B)
lakes were higher than in all areas except 1D. Me-
dian sodium concentrations in the Southern Blue
Ridge were approximately half those in Florida.
The population estimates for potassium showed
patterns among subregions similar to those for
sodium. The highest concentrations at Q4 in each
region were for Subregions 1D, 2D, and 3B. Differ-
ences in magnesium among subregions closely
paralleled those observed for calcium (Section
4.6.2). Median magnesium concentrations were
higher in Subregion 2D; the lowest Q1 values were
estimated for Subregions 1A and 2C. Within the
Northeast, magnesium was higher in the Poconos/
Catskilis (1B) and Southern New England (1D). The
lowest Q1 was in the Adirondacks (1 A) and the low-
est Q4 in Maine (1E). Magnesium concentrations in
the Upper Midwest were generally higher than
those in the Northeast. Within the Upper Midwest,
Q1 and median concentrations were lowest in Sub-
region 2C. Magnesium concentrations in Subre-
gion 2A were similar to the concentrations for
sodium and potassium, and had a smaller in-
terquintile range than other subregions.
4.8,4 Iron, Manganese and Total Aluminum
The lowest median (19.1 u-g L~1) and Q4 (68.8 u,g
L~1) values for iron were estimated for Florida (3B),
which also had the second lowest Q, value (3.7 jig
L~1) of any subregion (Table 4-29). The three
highest median and quintile values for iron were
estimated for Subregions 1 D, 2A and 3A. Within the
Northeast, Subregions 1A, 1C and 1E had the three
lowest estimates of median iron concentrations.
Higher iron concentrations were estimated for Sub-
region 1B and the highest for Subregion 1D. In the
Upper Midwest, iron concentrations in Northeast-
ern Minnesota (2A) were approximately twice that of
the other subregions at median values. In Region 2,
Q4 values for iron ranged from 196.6 JJQ L"1 in
Subregion 2D to 268.2 /jg L"1 in Subregion 2A.
Concentrations of iron in Subregion 3A were similar
to those in the Upper Midwest. As mentioned above,
Florida (3B) lakes had low levels of iron.
The highest concentration of manganese in the
Northeast occurred in the Poconos/Catskills (1B)
where the median and Q4 values were 28.1 and
121.4 u,g L"1, respectively. The lowest median and
Q4 values were observed in Maine (1E). The lowest
concentrations of manganese occurred in the Up-
per Midwest where all subregions except Northcen-
tral Wisconsin (2C) had median values of 0.0 u,g L"1.
The Southern Blue Ridge (3A) had the highest con-
centrations of manganese at all three values (Qi = 8.6,
median = 52.7 and Q4 = 156.5).
For the Northeast, total aluminum was lowest in the
Poconos/Catskills (1B) and highest at median and
Q4 values in the Adirondacks (1 A). Median, Q1 and
Q4 values for total aluminum concentration were
lower in the Upper Midwest than in the Northeast.
The lowest values for total aluminum were esti-
mated for Northcentral Wisconsin (2C) and the Up-
per Great Lakes Area (2D). The highest Q4 concen-
trations estimated in the Upper Midwest were
found in Subregions 2A and 2B, and these values
were similar to those in Subregions 1D and 1E.
Table 4-29.
Secondary Variables (Iron, Manganese and
Total Aluminum): First Quintiles (Q,), Medians
(M), and Fourth Quintiles (Q4), Eastern Lake
Survey-Phase I
Fe (fig L-1)
Mn (nig L-1) Al-total (p,g !_-')
M
-"1
M Q4
•"1
M
Q4
SUBREGION 1A 10.6 40.9 124.0 2.1 13.9 32.9 27.1 67.2 175.6
1B 18.3 69.3 127.9 0.0 28.1 121.4 16.8 36.1 62.0
1C 13.8 40.9 99.5 3.7 11.9 32.2 28.8 58.3 117.8
1D 29.4 84.0 248.3 0.0 19.0 84.1 20.3 48.1 103.6
1E 11.0 39.9 98.1 2.1 6.9 15.2 26.7 56.2 102.3
REGION
1 14.6 49.8 136.3 1.7 11.8 41.4 23.9 49.9 111.6
SUBREGION 2A 20.3 104.9 268.2 0.0 0.0 1.8 16.0 39.8 97.0
2B 13.8 49.9 201.2 0.0 0.0 20.4 12.5 30.9 107.6
2C 9.5 53.2 238.2 0.0 7.0 29.5 9.1 21.1 45.2
2D 2.1 44.0 196.6 0.0 0.0 9.1 8.0 19.7 48.1
REGION
8.2 55.3 208.2 0.0 0.0 17.1 9.1 22.9 61.7
SUBREGION 3A 29.7 83.4 255.1 8.6 52.7 156.5 39.0 81.7 185.6
3B 3.7 19.1 68.8 1.0 4.2 16.0 22.9 44.4 130.6
64
-------
The Southern Blue Ridge (3A) had the highest
estimated total aluminum values of any subregion.
Concentrations of total aluminum in Florida lakes
were similar to those in Subregions 1D and 1E, but
the Q4 value for Florida lakes was higher.
4.8.5 Other Secondary Variables
Estimated concentrations of silica were similar for
most areas of the Northeast, with the exception of
lower values in the Poconos/Catskills (1B) (Table 4-
30). The Upper Midwest had higher Q4 values than
the Northeast, but only slightly higher medians. An
exception to this was Northcentral Wisconsin (2C),
where concentrations at the Qi and median were
lower than in any other area in Region 1 or 2. The low
concentrations of silica were found in seepage lakes.
Concentrations of silica were high in the Southern
Blue Ridge (3A); Florida (3B) had the lowest concen-
trations of any area and, like Northcentral Wisconsin
(2C), most are seepage systems.
Dissolved inorganic carbon (DIG) levels were general-
ly low in the Northeast, with slightly higher concen-
trations in the Poconos/Catskills (1 B). In the Upper
Midwest, DIC was lowest at Qi and median values in
Northcentral Wisconsin (2C); it was highest in the
Upper Great Lakes Area (2D). The Q4 concentration
for Subregion 2D was the highest of any area
sampled. The Southern Blue Ridge (3B) had the
highest Qi value, but had a Q4 value that was close to
the regional value for the Northeast.
Concentrations of chloride showed patterns similar to
those for sodium (Section 4.8.3). Within the North-
east, chloride concentration was lowest in the
Adirondacks (1A) and highest in Southern New
England (1 D). In the Upper Midwest, concentrations
were generally low, although some sites in the Upper
Great Lakes Area (20) had higher values (Q4 = 1 06.2
//eq L"1). Chloride concentrations were high in Florida
(3B) (Q, = 1 34.7 /ueq L'1; Q4 = 537.3 A/eq L"1).
As expected, the patterns for conductance paralleled
those for major ionic constituents, particularly sodium
chloride and ANC. For the Northeast, conductance
was lowest in the Adirondacks (1 A) and Maine (1 E),
and highest in the Poconos/Catskills (1B) and
Southern New England (1D). In the Upper Midwest,
the lowest Qi and median values were observed in
Subregion 2C. The least interquintile difference was
observed in Subregion 2A, again related to the
relatively uniform water chemistry in Northeastern
Minnesota (2A) compared to other areas. High Q4
values of conductance were found in Subregions 2B
and 2D. Florida (3B) lakes had Qi and median values
similar to those in Southern New England (1 0), and
the highest Q4 of any area.
4.9 Characteristics of Special Interest
Lakes
During the ELS-I, 186 special interest lakes were
sampled (Section 2.2.5). Population estimates were
not calculated using data from these lakes because
they were not selected as part of the probability
sample. Consequently, the characteristics of these
lakes are summarized separately by region and pre-
sented in Table 4-31. Data for each special interest
lake are presented in Volume III.
In general, special interest lakes were larger and
had greater watershed areas compared to esti-
mated population medians for the lakes in the prob-
ability sample. For nearly all variables, median val-
ues for the special interest lakes were lower than
Table 4-30. Secondary Variables (Silica, Dissolved Inorganic Carbon, Chloride, Conductance and Bicarbonate):
(Q,), Medians (M), and Fourth Quintiles (Q4), Eastern Lake Survey -Phase 1.
Conductance
SiO2 (mg L~1) DIC (mg L~1) Cl~ (fieq L~1) (/js cm"1) i
SUBREGION 1A
IB
1C
10
1E
Q!
0.8
0.3
0.8
0.5
0.8
M
2.3
1.0
1.9
2.1
1.9
Q4
4.5
2.7
3.8
4.8
3.1
QI
0.6
1.5
0.8
0.9
1.0
M
1.6
4.1
1.6
2.2
1.8
04
3.3
7.1
4.6
4.9
3.4
QI
8.0
43.1
12.2
149.1
10.6
M
11.1
109.5
42.7
382.2
32.5
04
56.5
312.3
213.0
827.0
88.5
0-1
21.7
39.1
22.2
47.4
23.3
M
30.7
63.7
34.9
81.5
33.0
04
52.8
105.1
94.6
156.3
48.2
QI
7.3
91.8
40.2
30.8
57.4
: First Quintiles
HC03~
M
102.9
258.3
102.6
141.0
122.0
Q4
218.6
459.1
339.6
358.4
264.2
REGION
1
0.6 1.9 3.8 0.9 2.1
5.0
SUBREGION 2A
2B
2C
0.9
0.3
0.1
20 0.3
2.6
2.3
0.6
2.4
5.4
6.1
6.6 0.5
9.0
1.3 2.4
1.0 4.5
1.6
1.7
4.7
14.8
7.3
9.5 22.1
12.1
6.3
6.1
6.5
11.9
59.6 293.6 25.3 43.3 95.4 40.7 137.4 350.8
8.8 14.1 22.3 30.4 52.6
10.2 22.5 20.5 47.2 132.9
11.3 36.7 13.4 22.5 73.2
22.0 106.2 24.7 91.6 198.7
66.0 143.7 342.2
29.2 274.7 1158.5
9.5 72.2 577.5
70.8 762.5 1900.0
REGION
0.3 2.3 6.8 1.2 4.4 16.1
7.6 16.8 65.8 20.7 44.3 142.8 51.7 287.4 1279.6
SUBREGION 3A 6.8 9.0 13.6 2.3 3.7 5.3 28.7 65.7 94.9 20.9 38.5 63.9 136.6 236.5 382.3
3B 0.1 0.3 1.2 0.6 1.7 10.2 134.7 222.3 537.3 37.1 90.8 197.6 4.0 76.0 651.8
65
-------
Table 4-31. Sample Statistics for Special Interest Lakes by Region: Minima (MIN), Medians (MED) and Maxima (MAX), Eastern
Lake Survey-Phara I
Region 1 (n=115) Region 2 (n=52) Subregion 3A (n=10) Subregion 38 (n=9)
Variable (Units)
Lake Area (ha)
Watershed Area (ha)
Site Depth (m)
ANC (n,eq L~1)
pH, closed
Sulfate (ixeq Lr1)
Calcium (^eq L~1)
Al-extractable (jig L~1)
DOC (mg L'1)
Conductance (jxS cm~1)
Chloride (|j,eq L"1)
Sodium (|ieq L~1)
Potassium (^.eq L~1)
Magnesium (p-eq L~1)
Silica (mg L"1)
DIG, open (mg L"1)
Al-total dig L"1)
Nitrate (p-eq L~1)
Ammonium (n,eq L~1)
P, Total ((ig L-1)
True Color (PCU)
Seech i Disk
Transparency (m)
MIN
0.9
5
0.6
-63.2
4.2
53.5
25.4
0.8
0.1
13.8
6.1
3.7
2.1
8.4
0.1
0.2
3
0
0
0
0
0.6
MED
21.2
295
5.2
42.4
6.4
108.8
85.0
11.8
3.1
23.6
11.1
30.2
8.3
31.4
1.9
0.9
105.8
0.4
1.3
5.1
20
3
MAX
597.7
13937
31.1
600.4
7.6
348.1
536.9
291.1
14.3
164.6
1063.9
1038.8
35.1
310.1
11.3
6.9
685
18.4
14.9
31
150
10.8
MIN
3.2
13
3
-18.5
4.9
4.6
26.4
0
1.1
8.1
5.1
6.5
6.9
19.7
0
0.2
0
0
0
0
5
0.9
MED
44.6
228
6.6
69.5
6.4
72.4
85.6
3
4.2
20.1
9.4
23.3
12.4
55.5
0.2
1.2
22.5
0.5
1.1
12.8
18.8
3.1
MAX
338.2
24577
24.7
990.6
7.7
148.4
632.7
80.5
21.3
98.1
254.2
204.9
23.8
422.8
18.5
11.8
199.5
9.5
9.4
95
135
5.5
MIN
2
109
3
16.7
5.9
16.1
14.5
1
0.7
7.8
14.8
37.4
7.2
9.0
3.3
0.7
31
0.6
0.3
2
10
1.6
MED
14.7
1064.5
6.2
128.2
6.8
27.3
79.8
2.5
1.6
19.6
24.1
51.1
15.5
36.6
6.0
1.9
53
1.5
1.9
5.8
27.5
2.3
MAX
2861.9
120474
60
498.9
7.2
110.6
286.9
8
4.1
75.8
85.5
97.0
34.5
244.3
9.9
6.6
380
34.8
40.6
11
55
4.4
MIN
16.6
67
2.7
-19.4
5.0
25.4
26.5
0
0.5
14.2
52.9
48.9
2.3
21.8
0.1
0.2
13
0
0
2
5
1.5
MED
81.4
355
5.5
-11.6
5.2
67.2
41.2
2
1.7
30.2
141.6
127.7
3.6
44.2
1.0
0.4
39
0.4
0.9
4.5
10
3.7
MAX
722.7
5541
17.7
131.6
7.0
699.6
376.1
64
6.7
169
558.6
582.9
91.8
358.1
2.8
1.4
200
5.6
1.6
23
70
7.0
the calculated population medians, with the excep-
tion of higher concentrations of sulfate in the Upper
Midwest, total aluminum in the Northeast, and sil-
ica in Subregion 3B. There were substantial differ-
ences between the sample medians of special inter-
est lakes and the estimated population medians in
the Northeast for sulfate, silica, nitrate and ammo-
nium.
There are three general categories of special inter-
est lakes: (1) the EPA-sponsored Long-Term Moni-
toring (LTM) lakes; (2) the lakes recommended by
the National Research Council (NRC) as having high
quality historical water chemistry data; and
(3) other lakes recommended by State resource
management agencies. The LTM lakes were origi-
nally selected in large part, because they had low
concentrations of ANC and it was expected that
they would have low ionic concentrations.
It was expected that the special interest lakes rec-
ommended by the NRC and State agencies would
also have low ANC based on their location in areas
previously identified as having populations of low
alkalinity lakes (e.g., Adirondacks, New Hampshire,
and Northcentral Wisconsin). The sample data for
the special interest lakes confirm that these lakes
have lower ANC than the lakes selected in the prob-
ability sample.
66
-------
Section 5
Results and Discussions of Associations Among Variables
The primary objectives of the Eastern Lake Sur-
vey-Phase I (ELS-I) were to provide estimates of
low ANC and low pH lakes in the regions of interest
and to characterize the populations of lakes for fur-
ther study in Phases II and III. The Survey was de-
signed to meet these primary objectives, and as
such, the results presented in Section 4 are appro-
priately presented on a univariate basis. It is also
possible to explore relationships among variables
using the ELS-I data base.
The purpose of this section (Section 5) of the report
is to present briefly an examination of some bivari-
ate and multivariate relationships that may be perti-
nent to addressing the role of certain variables in
lake acidification. However, the focus of these data
presentations, as in Section 4, continues to be
largely descriptive. The results are presented in a
manner that highlights differences and similarities
among subregions, but which avoids formal tests
of statistical significance among or within subre-
gions. Investigators who desire to use these data to
develop predictive models on populations of lakes
need to consider how the identification of lake pop-
ulations may influence model results. In particular,
those variables that may be related to alkalinity
map classes may need to be treated on a stratum-
by-stratum basis (Section 2.2.1).
When evaluating relationships among lakes it is
also important to consider that weighting factors
differ among subregions. Thus, multivariate mod-
els based upon sample data may incorrectly de-
scribe a relationship for a population of lakes. Un-
like the population estimates presented in
Section 4, the relationships presented in Section 5
include data from special interest lakes and lakes
>2000 ha unless noted otherwise.
5.1 Relationships of pH and ANC
5.1.1 Comparison of pH Measurements
The pH of each lakewater sample was measured
five different times on three separate aliquots (Sec-
tion 2.7). In situ pH was measured using the Hydro-
lab unit. At the field laboratory, pH was measured
on a sample from a syringe sealed from the atmos-
phere. At the analytical laboratory, subsamples
from a single aliquot were prepared for pH mea-
surements at the start of separate titrations for ANC
and BNC and a separate subsample equilibrated
with the atmosphere. Pair-wise comparisons of
these five measurements were made using linear
regressions. The results of the regression analyses
are presented in Table 5-1, which shows that the
regression equations explained 87 to 99 percent of
the total variance. Most slopes were significantly
different from 1 (p <0.05) indicating that pH mea-
surements were not identical; however, they were
comparable and could be calibrated to one another.
Table 5-1. Comparison of pH Measurements for Regular
and Special Interest Lakes, Eastern Lake Survey-
Phase I
pH Pair
Dependent/Independent Variable n Slope Intercept r2
Closed systema/ln situb
Closed system/Open system
(ANC)C
Closed system/Open system
(BNC)d
Closed system/Air-equilibratede
In situ/Open system (ANC)
In situ/Open system (BNC)
In situ/Air-equilibrated
Air-equilibrated/Open system
(ANC)
Air-equilibrated/Open system
(BNC)
Open system (ANQ/Open sys-
tem (BNC)
1779
1793
1793
1781
1779
1779
1767
1781
1781
1794
0.938
0.989
0.979
0.811
1.002
0.991
0.827
1.166
1.148
0.985
0.535
0.101
0.123
0.883
-0.116
-0.087
0.642
-0.617
-0.546
0.057
0.943
0.942
0.942
0.905
0.901
0.899
0.875
0.954
0.942
0.989
3Syringe sample, unexposed to the atmosphere.
bin situ pH is the value measured in the lake using the Hydrolab.
cLaboratory measurement at initiation of the ANC titration.
dLaboratory measurement at initiation of the BNC titration.
"Laboratory measurement of an aliquot after bubbling with 300
ppm CO2 for 20 minutes. However, complete equilibration with
atmospheric C02 may not have been obtained.
The field laboratory pH measurement was used in
interpreting ELS-I data (Section 2.6.4, closed sys-
tem pH). The field laboratory measurement was
performed within 16 hours of collection on a sam-
ple that was unexposed to the atmosphere during
collection and analysis, minimizing the problem of
a change in pH over time caused by CO2 exchange
during storage. The measurement in the field labo-
ratory was made using more sophisticated equip-
67
-------
ment than was used for the in situ measurement,
reducing the problem of instrument instability. The
closed system pH represented the best single mea-
surement of in situ pH conditions, and hence was
• used for all population estimates.
The in situ measurement was not selected for use in
interpretations because the precision was lower
than the field laboratory measurement. The three
analytical laboratory pH measurements were not
used for data interpretation because the samples
either gained or lost C02 and therefore differed from
the value observed in the lake. As shown in Figure
5-1, for closed and open system pH, much of the
variation observed between different measurements
occured at pH values greater than six. The r2 for pH
values <6.0 was 0.96; for pH >6.0, r2 was 0.83. This
pH-dependent variation is consistent with changes
caused by C02 exchange between the sample and the
atmosphere. This is supported by comparing closed
system pH to air-equilibrated pH. These measure-
ments were made using identical instrumentation yet
the regression equation did not have a slope of 1.
5.1.2 Relationship Between pH and ANC
Figures 5-2 through 5-5 depict the relationship be-
tween closed system pH and ANC for Regions 1 and
2, and Subregions 3A and 3B. Also plotted on each
figure is the curve showing the expected relation-
ship between pH and ANC assuming the lakes are
Figure 5-1. Closed system pH versus initial open system pH for probability sample lakes and special interest lakes in all regions.
Eastern Lake Survey-Phase I.
8.5 —
7.5 —
_ 6.5-
•o
-------
Figure 6-2. pH (closed system) versus ANC (/ueq L~1)
are shown as x's and darkwater (>30
10"35atmCO2.)
PCLI)
8.5-
7.5 —
6.5-
o
o^
o- 5.5
4.5 -
3.5.
For Region I. Eastern Lake Survey-Phase I. (Clearwater lakes (<30 PCU)
as triangles. The theoretical curve shown was calculated for 2S°C and
x Clearwater Lakes
A Darkwater Lakes
-200
I
200
I
400
ANC (/ueq L'1)
I
600
800
1000
in equilibrium with atmospheric C02(pCO2= 10"3'5
atm at 25°C; Butler 1982). Clearwater (color $30
PCU) and darkwater lakes (color >30 PCU) are
represented by X's and triangles, respectively.
Both the lakewater temperature and the partial
pressure of COa influence the solubiltiy of C02 in
water (Stumm and Morgan 1981) and the position of
the theoretical curve (Butler 1982). The water tem-
perature assumed for calculation of the theoretical
curves in Figures 5-2 through 5-6 was 25°C. The field
laboratory pH samples, however, were taken from
lakes at temperatures below 25°C. The samples were
allowed to equilibrate to ambient air temprature but
were not allowed to equilibrate with the atmosphere
prior to pH measurement. This caused the samples at
laboratory temperature to be highly supersaturated
with CC*2, and resulted in lower pH readings than
expected compared to the theoretical curve. One
remedy for this problem is to recalculate the theoreti-
cal relationship for each region, using the average
lake temperature for that region. When this was
done, the theoretical curves were still above most
data points but not as far above. The theoretical
curves shown in the figures, however, were identi-
cally calculated for a temperature of 25°C for ease in
making comparisons among subregions.
-------
Figure 6-3. pH (closed system) versus ANC (/neq L'1) for Region 2. Eastern Lake Survey-Phase I. (Clearwater lakes (<30 PCU)'
are shown as x's and darkwater (>30 PCU) as triangles. The theoretical curve shown was calculated for 25°C and
10"35atmCO2.)
8.5-
7.5-
E
i)
Q.
5.5-
4.5-
3.5-
** | .«4- f
* » » *"„«» » *
1 «..*..-.•-. **
x Clearwater Lakes
A Darkwater Lakes
-200
I
200
I
400
ANC (/L
-------
Figure 5-4. pH (closed system) versus ANC (/»q L"1) for Subregion 3A, Eastern Lake Survey-Phase I. (Clearwater lakes (<30
PCU) are shown as x's and darkwater (>30 PCU) as triangles. The theoretical curve shown was calculated for 25°C
and 10~35atmCO2.)
8.5-
7.5H
E
03
6'5"
a
5.5-
4.5-
3.5-
x Clearwater Lakes
A Darkwater Lakes
-200
I
0
200
400
ANC (peg L"1)
600
800
1000
The data were much closer to the theoretical curve
at lower values of pH and ANC, indicating that the
relative importance of supersaturation with CO2 in
these lakes increases with increasing pH and ANC.
Lakes with high ANC are likely to be more biologi-
cally productive than low ANC lakes (Ryder 1964).
Biological activity can strongly affect concentra-
tions of C02 in lakewater (Moss 1973) often increas-
ing pC02 values in the water column.
As the theoretical relationship considers strictly bi-
carbonate alkalinity, the presence of other con-
stituents in lake water may cause measured sam-
ples to deviate from the predicted relationship. The
presence of weak acids may result in a lower pH
and higher ANC than those predicted by the theo-
retical relationship, and thus, samples containing
high concentrations of weak acids may plot below
the theoretical line. In both Regions 1 and 2, at low
ANC values, Clearwater lakes were closer to the the-
oretical curve than were darkwater lakes. This may
be explained by the observation that darkwater
lakes have higher DOC (Section 5.2.6.1), and there-
fore higher weak acid concentrations. Weak acids are
more likely to affect the ANC versus pH curve in this
region of the curvethan at higher pH and ANC values
(Small and Sutton 1986).
71
-------
Figure 5-5.
pH (closed system) versus ANC (/mq L~1) for Subregion 3B, Eastern Lake Survey-Phase I. (Clearwater lakes (<30
PCU) are shown as x's and darkwater (>30 PCU) as triangles. The theoretical curve shown was calculated for 25°C
and 10~35atmCO2.)
8.5-
7.5-
-------
Figure 5-6. pH (air-equilibrated) versus ANC (jueq L~1) for Region 1. Eastern Lake Survey-Phase I. (Clearwater lakes (<30 PCU)
are shown as x's and darkwater (>30 PCU) as triangles. The theoretical curve shown was calculated for 25°C and
10"35atmCO2.)
8.5-
7.5-
jo 6.5-
cr
ui
a.
5.5-
4.5-
3.5-
x Clearwater Lakes
A Darkwater Lakes
-200
I
0
200
T
400
ANC (/jeq L"
600
r
800
1000
less than 10 u,eq L~1, "endangered" if ANC was 10
to 50 (ieq L~1, or "satisfactory" if ANC was greater
than 50 (xeq L~1; the same terminology was applied
to lakes with pH less than 5.0,5.0 to 6.0, and greater
than 6.0, respectively. Because of the widespread
use of sensitivity indices, it is inevitable that the
Survey results will be discussed by others in these
terms. Table 5-2 presents Survey results using the
above indices.
Some authors have indicated that any single vari-
•able is a poor measure of lake sensitivity to acidic
deposition and proposed multiple factor indices
(Table 5-2). Zimmerman and Harvey (1979), for ex-
ample, considered a lake to be "sensitive" when
ANC was less than 300 jxeq L"1, pH was less than
6.3, and conductance was less than 30 n-S crrr1.
The estimates of the number of sensitive lakes are
highly dependent on both the variables selected for
sensitivity indices and the reference values chosen
for each variable.
No simple classification criteria have been pro-
posed to account for the contribution of organic
acids to lake acidity or to represent the potential
73
-------
Table 5-2. Eastern Lake Survey- Phase I (ELS-I) Population Estimates for Selected Literature Definitions of Sensitivity. (ELS-I
Reference Values Are Included for Comparison Only and Are Not Proposed as Sensitivity Classes.)
Number of Lakes s2000 ha by Region and Subregion
Sensitivity Classification
SINGLE FACTOR
Critical6'0
Endangered6
ELS Reference0
Satisfactory6
ELS Reference0
Critical6
Endangered6
ELS Reference0
Satisfactory6
Extreme sensitivity"1
Moderate to high sensitivity*1
Sensitive6
Susceptible or senstive°-f
Sensitives
Definition3
PH
pH
PH
pH
ANC
ANC
ANC
ANC
ANC
ANC
ANC
ANC
ANC
ANC
==5
5-6
±=6
>6
sO
==10
10-50
±=50
>50
<40
40-200
±=100
±=200
±=140.
1
240
676
916
6180
326
537
827
1364
5732
1203
3055
2482
4258
3242
1A 1B
128 12
215 103
343 116
947 1363
138 78
232 92
227 102
459 194
831 1285
404 181
505 391
607 288
909 572
756 352
1C
25
166
191
1292
35
75
187
262
1221
244
758
652
1002
848
1D
66
126
192
1126
66
113
172
284
1034
258
497
420
755
552
1E 2
8 130
65 688
74 818
1452 7683
8 148
25 340
140 972
165 1312
1361 7189
116 946
903 2572
516 2174
1020 3518
734 2855
2A
0
20
20
1437
0
0
60
60
1396
28
802
309
830
475
2B
99
87
185
865
102
136
62
198
852
192
246
299
438
342
2C 2D
31 0
380 202
41 1 202
1069 4312
45 0
195 9
418 432
612 441
868 4073
548 178
291 1233
738 828
839 1411
781 1258
3A 3B
0 259
1 427
1 687
257 1412
0 463
0 504
4 238
4 742
254 1357
2 690
86 466
28 1073
88 1156
43 1075
MULTIPLE FACTOR
Sensitiveh
pH ±=6.3,
conductance =£30,
ANC <300,
858 408 102 229 27 92 1341 93 250 601 397 3 260
ESTIMATED TOTAL
NUMBER OF LAKES, ELS-1
7096 1290 1479 1483 1318 1526 8501 1457 1050 1480 4515 258 2098
aUnits for pH are standard units, ANC (|j,eq L~1), conductance
bPfeiffer and Festa 1980.
"Reference values for Eastern Lake Survey-Phase I.
dAnon, 1981.
eOmernik, unpublished maps.
'Altshuller and MacBean 1980; Hendrey et al. 1980.
flQalloway 1984.
"Zimmerman and Harvey 1979.
cm-1), Ca (mg L~1).
sensitivity of darkwater lakes. Organic ligands can
complex aluminum, thereby mitigating its toxic ef-
fects on fish (Driscoll et al. 1980). However, the re-
sponse of darkwater systems to acidification is still
poorly understood (Gorham et al. 1984). Some ad-
ditional nonchemical factors are also considered
important in determining lake susceptibility to acid-
ification. These factors include: hydrology (Eilers et
al. 1983), weathering rates (Schnoor et al. 1983),
depth of overburden (Chen et al. 1984), lake sedi-
ments (Baker et al. 1985; Cook and Schindler 1983),
and land use (Krug and Frink 1983).
In conclusion, any simple definition of sensitivity
has inherent limitations that must be recognized.
Comparisons of population estimates based on
proposed indices demonstrate that conclusions are
likely to vary depending on the index used. Until
greater progress is made in understanding the
processes that control acidification, it is difficult to
justify the use of any of these indices except for the
purpose of comparing the chemistry of lakes.
5.2 Selected Associations Among
Chemical Variables
5.2.7 Introduction
In addition to addressing the primary objectives of
the ELS-I, the Survey data also provided an oppor-
tunity to evaluate relationships among variables.
Along with pH and ANC, the samples were also
analyzed for major anions, major base cations, true
color, DOC, silica, aluminum, iron, manganese, and
conductivity (Sections 2.5 and 2.6). It is beyond the
scope of this report to present an indepth analysis
of the relationships among all variables measured
during the Survey. However, described here is an
overview of some important associations germane
to lake acidification processes and hypotheses. As-
74
-------
sociations do not prove that a causal relationship
among variables exists, nor is absence of an associ-
ation proof that a causal relationship does not exist,
but they are useful in formulating or examining rel-
evant hypotheses. Observations that are repre-
sented by substituted values (Section 3.4) have
been excluded from these analyses.
5.2.2 Sulfate
Many investigators have attempted to draw infer-
ences regarding the extent of acidification by com-
paring pH, [Ca+2 + Mg+2] and S04~2 concentrations
(Henriksen 1980). Efforts have also been made to
correct the measured sulfate values for sea-salt
contributions. This correction permits comparison
of "excess" or sea-salt corrected lakewater sulfate
to the chemistry of deposition.
The most common procedure for calculating the
wet "excess" or sea-salt corrected sulfate has been
to relate the S04 :CI~ or S04 :Mg z in sea water to
the ratio in lake water. Lake sulfate values are cor-
rected by subtracting the equivalent concentration
of sulfate possibly associated with chloride or mag-
nesium. This calculation has been useful in small
geographic areas with relatively uniform geologic
bedrock and lake types, such as Norway (Henriksen
1980).
Several points should be considered when propos-
ing to make sea-salt corrections to sulfate concen-
trations:
(1) salt concentrations in rain water derived from
oceanic sources may not occur in the same
ratios as in sea water (Kroopnick 1977);
(2) sulfate is not a conservative ion in many
watersheds and lakes (Johnson and Reuss
1984) and lake sediments (Cook and Schin-
dler 1983);
(3) use of the sea-salt corrections assumes that
the ion paired with SO4~2 is derived from sea-
salt; i.e., there are no significant sources of
Cr or Mg+2 in the watershed. In cases where
natural sources of the paired ions occur, the.
"corrections" may be misleading; and
(4) the presence of anthropogenic sources of
chloride (e.g., road salt), invalidate the correc-
tion (Wright 1983).
For these reasons, it was not useful to report sulfate
results in the sea-salt corrected form. For example,
when sea-salt corrections were applied to several
areas, many negative S04~2 values resulted. Conse-
quently, all sulfate data presented in this report are
reported as actual, measured concentrations.
Dickson (1975,1978) proposed that elevated sulfate
concentrations in lakes should also be strongly re-
lated to pH and that the highest S04~2 concentra-
tions should correspond to low pH values. This
assumes that S04~2 is conservative on its path from
atmospheric deposition through the watershed and
into lakes and that within-lake sulfur cycles are not
significant sinks. The results of the Survey do not
provide evidence to support this hypothesis. No
correlation between pH and SO*"2 was observed in
the three subregions with large numbers of low pH
lakes: the Adirondacks (1A), the Upper Peninsula of
Michigan (2B), and Florida (3B) (Figures 5-7 through
5-9).
It is noteworthy that in the Adirondacks (1A) the
concentrations of sulfate described in Figure 5-7
were the highest of these three subregions, approx-
imately double those in the Upper Peninsula of
Michigan (2B). This is consistent with the hypothe-
sis that the northeastern lakes have been chemi-
cally altered by high levels of atmospheric sulfate
deposition, although it does not exclude the possi-
bility that watershed sources contribute dispropor-
tionate loadings of S04~2among subregions.
5.2.3 Extractable Aluminum
5.2.3.1 Background
The analysis of chemical forms of aluminum (Al) in
natural systems is fraught with sampling and ana-
lytical problems (Barnes 1975, Driscoll 1984). In
cases where information about ionic forms of Al is
desired, analytical differentiation of colloidal, sus-
pended forms from dissolved forms is possible.
However, commonly used analytical procedures
cannot differentiate among the Al in monomeric
dissolved form and any polymeric, colloidal, or
crystalline, forms in natural waters. Even filtration
through 0.1 fxm pore-size filters is not adequate for
separating and distinguishing dissolved and col-
loidal Al (Barnes 1975).
The method used in the ELS-I to measure dissolved
and readily reactive Al species (extractable alu-
minum) is based on the operationally defined pro-
cedure developed by Barnes (1975). In this method,
Al is complexed with 8-hydroxyquinoline and the
complex quickly extracted into methyl isobutyl ke-
tone (MIBK) solvent. Equilibrium partitioning of the
Al-oxine complex is accomplished in 10 seconds at
20°C if the mixture is shaken vigorously; complete
extraction of monomeric Al occurs under these
conditions (Barnes 1975). This extractable Al can be
referred to as dissolved monomeric Al and occurs
as aquo aluminum (Al+3), as well as hydroxide, flu-
oride, sulfate, and organic complexes of Al (Rober-
son and Hem 1969).
75
-------
Figure 5-7. pH (closed system) versus sulfate (jueq L'1) for probability sample lakes in Subregion 1 A. Eastern Lake Survey-Phase I.
8.5-
7.5-
•g 6.5-
I
w'
•O
CO
^
O
^ 5.5H
4.5-
3.5-
% '
40
80
120
160
200
Sulfate (jueq L~
Many of the methods used to measure low concen-
trations of Al in natural waters involve Al complex-
ation and subsequent analysis of the resulting com-
plex by spectrophotometry or fluorometry. The
presence of other complexing metals (Fe and Mn)
often limits the use of such methods. Also, the
specificity of certain methods depends on the kinet-
ics of the complexation reaction. If the time allowed
for complexation is too long, complexation with
microcrystalline or adsorbed Al may result in an
overestimation of soluble monomeric Al. There-
fore, such methods are time-dependent, opera-
tionally defined analyses.
The use of filtration to separate paniculate forms of
Al from the soluble species has received scrutiny
(Barnes 1975). Changes in pH can result from C02
degassing which is enhanced by the turbulence of
filtration. Hydrolysis can occur rapidly with pH
changes, and hydrolyzed Al adsorbs more readily to
filters. In solutions that contain concentrations near
or above Al saturation, filtration may cause loss of Al
from the soluble phase through precipitation or
adsorption. Adsorption of soluble Al to particles
present on the filter or to the filter itself may
significantly decrease soluble Al concentrations.
Chemical interferences of concern in the ELS-I data
include samples containing high concentrations of
Fe and/or Mn. Incomplete recovery of monomeric
Al occurs when Fe or Mn exceeds 1 mg L~1. A total
76
-------
Figure 5-8. pH (closed system) versus sulfate (/-
CO
o
i
5.5-
4.5-
3.5-
* *• .
•
« I
« • «
• • • * •
• •
• »
.' • V -..V
40
l
80
120
160
200
Sulfate (fjeq L"1)
of 14 lakes had Fe concentrations in excess of 1 mg
L"1, nine of which were from Region 2. Only three
lakes had Mn concentrations above 1 mg L~1.
Therefore, chemical interference from Fe and Mn
likely is insignificant in this study.
5.2.3.2 Associations Between Extractable
Aluminum and Other Variables
Extractable aluminum concentration is expected to
increase with decreasing pH (Cronan and Schofield
1979; Driscoll et at. 1980). Log [extractable Al] was
regressed on pH for the Northeast, Upper Midwest,
Southern Blue Ridge and Florida (Table 5-3). Fig-
ures 5-10 through 5-13 show untransformed values
of extractable Al versus pH for each of these areas.
Table 5-3.
Region
Regression Statistics for Log [Extractable Al]
(Molar, Dependent) versus pH (Independent) for
All Regions, Eastern Lake Survey-Phase I
Slope
Intercept
1
2
3A
3B
842
590
106
124
-0.457
-0.350
-0.277
-0.120
-3.52
-4.27
-5.18
-5.50
0.41
0.27
0.12
0.11
Extractable Al concentrations are generally greater at
low pH values. The relationships were not consistent
among the areas: however, the Northeast showed
greater values of extractable Al and a steeper
increase with decreasing pH than other areas (Table
5-3, Figure 5-10). Lakes in Florida (3B) showed the
least increase in Al at low pH values (pH <5.0).
77
-------
Figure 5-9. pH (closed system) versus sulfate (//eq L~1) for probability sample lakes in Subregion 3B, Eastern Lake Survey-Phase I.
8.5-
7.5-
6.5-
W
T3
0)
0>
O
o
Q.
5.5-
4.5-
3.5-
I
40
I
80
120
160
200
Sulfate (//eq L"
Darkwater lakes and clearwater lakes had similar
relationships between extractable Al and pH in the
Northeast. In the Upper Midwest, darkwater lakes
showed higher extractable Al values for a given pH
than did clearwater systems (Figure 5-11). Ex-
tractable Al is apparently compiexed by organic
materials and then extracted by the MIBK. Almost
all clearwater lakes with pH ^5.0 and high ex-
tractable Al occurred in the Northeast. Most of
these lakes were found in the Adirondacks (1 A).
Caution should be used in interpreting the informa-
tion, since extractable Al analyses are highly vari-
able. The system decision limits, values calculated
as the 95th percentile of field blank analyses, were
8 |xg L~1 for extractable Al and 30 jig L~1 for total Al.
If the system decision limit were used to exclude
values near zero, approximately half the data from
Regions 1 and 2, and Subregion 36, and more than
90 percent of the data from Subregion 3A would be
excluded from Figures 5-1 Othrough 5-13. Gilliomet
al. (1984) suggest that data below detection limits
are still useful in detecting data trends. As seen in
Figures 5-10 through 5-13 the data with extractable
Al values below the system decision limit (<8 fxg
L~1) demonstrate a similar relationship with pH as
those data above the decision limit.
Population estimates of median values of the ratio
of extractable Al to total Al ranged from 0.03 in
Subregion 3Ato 0.22 in Region 2 and Subregion 3B.
The range of values for this ratio for all data was
78
-------
Figure 5-10. Extractable Al (jug L'1) versus pH (closed system) for Region 1, Eastern Lake Survey-Phase I. (Clearwater lakes
(<30 PCU) are shown as x's and darkwater (<30 PCU) as triangles.)
200-
160-
120-
s
LU
E
3
C
E 80-
3
40-
"x x
x Clearwater Lakes
A Darkwater Lakes
** ww V K • *
x» *X A A
X X 4 4*x V V.A A
A * X 4 A . Arf 4
t « • „ X X"X A4 »-*AU*^ A A X 4
• • ••••^attiSc
3.5
4.5
I
5.5
6.5
pH (Closed System)
\
8.5
from 0 to >1, as measured extractable Al exceeded
total Al at some sites. This ratio was plotted against
DOC for all data and for each subregion but no
significant relationships were found.
5.2.4 ANC versus Base Cations
Calcium and magnesium are the primary base
cations and bicarbonate is the dominant anion in
surface waters. Calcium and magnesium comprise
nearly eighty percent of the cations in freshwater
on a world average (Livingstone 1963). In many
watersheds calcium and magnesium are the pri-
mary weathering products. An assumption of car-
bonic acid weathering is that divalent cations and
bicarbonate are produced in equivalent amounts.
Consequently, the sum of calcium and magnesium
is related directly to alkalinity4 resulting from car-
bonic acid weathering reactions. Departures from
this linear relationship, to a slope of less than about
one and to a negative intercept, have been used as
an indication of acidification (Aimer et al. 1978). For
thirteen reference areas with low sulfate deposi-
tion, Henriksen (1980) found a relationship between
non-marine [CA+2 + Mg+2] and alkalinity where
ALK = 14 + 0.93 [Ca* + Mg*], where Ca* and Mg*
represent sea-salt corrected concentrations.
4For purposes of this analysis, ANC is treated interchangeably with alka-
linity and bicarbonate. Distinction between these terms can be found in
Kramer (1984).
79
-------
Figure 5-11. Extractable Al (//g L 1) versus pH (closed system) for Region 2, Eastern Lake Survey-Phase I. (Clearwater lakes
(<30 PCU) are shown as x's and darkwater (<30 PCU) as triangles.)
200-
160-
120-
I
E
3
C
80-
40-
0 ~
x Clearwater Lakes
A Darkwater Lakes
X X *
3.5
4.5
x
*x *
6.5
pH (Closed System)
7.5
8.5
Henriksen (1980) developed a predictor nomograph
based on the relationships between ANC and
[Ca + Mg] using regressions from Norwegian data
between rainwater sulfate and lakewater sulfate,
and those between lakewater pH and lakewater
[Ca+2 + Mg+2]. The nomograph has been used to
infer acidification from the deficit of ALK relative to
concentrations of [Ca+2 + Mg+2]. It has also been
used to estimate steady-state alkalinity assuming
an increase in the acidity of precipitation. One of the
components of the nomograph is the regression of
[Ca+2 + Mg+2I on ALK. When forced through the
origin, this yields ALK = 0.91 [CA* + Mg*].
Examples of the difficulties encountered when ap-
plying the nomograph without modification to
North America are found in Haines and Akielaszek
(1983) and Church and Galloway (1984), who at-
tempted to apply the model to New England and
the Adirondacks, respectively. One of the diffi-
culties encountered by Wright (1983) was in esti-
mating background sulfate concentrations for each
region of North America for use in the empirical
model of Henriksen.
It was not possible to apply the predictor
nomograph without calibration to the results of the
Survey because many of the assumptions neces-
sary for its application were not met. For example,
lakes with significant organic anion concentrations
(usually indicated by high DOC) have anion deficits,
and may also have increased concentrations of
base cations due to chelation by organic complexes
(Henriksen 1979, Oliver et al. 1983). Consequently,
80
-------
Figure 5-12. Extractable Al (yug L"1) versus pH (closed system) for Subregion 3A, Eastern Lake Survey-Phase I. (Clearwater lakes,
(<30 PCU) are shown as x's and darkwater (<30 PCU) as triangles.)
200-
160-
120-
CO
s
UJ
E
3
80-
40-
x Clearwater Lakes
A Darkwater Lakes
3.5
4.5
5.5 6.5
pH (Closed System)
7.5
8.5
high DOC lakes have been excluded from previous
applications of the nomograph. The leaching of
base cations from soils increases under conditions
of mineral acid weathering, but the increase can
only be approximated (Henriksen 1982; Wright
1983). Differences in bedrock, geology, hydrology,
in-lake processes and sources of sulfur and other
ions vary from area to area, likely requiring recali-
bration to each area of application.
None of the areas sampled in ELS-I yielded signifi-
cant relationships between sulfate concentrations
and [Ca+2 + Mg+2]. A large degree of scatter in this
relationship was observed in all subregions, with
high sulfate values found at both low and high con-
centrations of base cations. Wright (1983) also ob-
served no relationship between S042 and
[Ca+2 + Mg+2] for three reference areas in North
America using sea-salt corrections and concluded
that background sulfate concentrations were re-
gional in nature. Background excess sulfate was
estimated at 26 to 41 jieq L~1, with somewhat
higher concentrations in the Upper Midwest
(Wright 1983).
The relationship between base cations and ANC
(such as ANC versus [Ca+2 + Mg+2]) is useful in il-
lustrating differences between weathering prod-
ucts and regimes in the subregions sampled. A
slight deficit of ANC relative to base cations
'[Ca+2 + Mg+2] (intercept between -7 and -27 jjieq
L~1) has been observed in several reference areas
81
-------
Figure 5-13. Extractable Al !/ug L~1) versus pH (closed system) for Subregion 38, Eastern Lake Survey-Phase I. (Clearwater lakes
(<30 PCU) are shown as x's and darkwater (<30 PCU) as triangles.)
200-
160-
120-
ID
S
m
E
3
80-
40-
x Clearwater Lakes
A Darkwater Lakes
X Xft, „
XX
XX M
3.5
I
4.5
I
5.5
6.5
pH (Closed System)
7.5
I
8.5
where weathering is due principally to carbonic
acid (Henriksen 1980; Liaw 1982; Wright 1983). In-
tercepts more negative than —27 |xeq L~1 have been
regarded as evidence of some mineral acid or non-
carbonic acid weathering (Henriksen 1980). Thus,
deficits >-27 |j,eq L~1 should be proportional to the
concentration of sulfate (Kramer and Tessier 1982).
Henriksen (1980) calculated a slope of 0.93 for lakes
in areas not receiving acidic deposition. Values of
slope <0.9 have been attributed to increases in
base cation weathering rates, also due to the pres-
ence of mineral acids. In this case, less ANC is
present than would be expected on the basis of the
concentrations of [Ca+2 + Mg+2].
BJ
Table 5-4 gives the regression statistics for the rela-
tionship of ANC versus [Ca+2 + Mg+2] by subregion
for ELS-I. Statistics are given for ANC as the de-
pendent variable, with an upper limit of ANC and
[Ca+2 + Mg+2] established as <200 (xeq Lr1. In all
cases, comparisons between all lakes and clear-
water lakes are given. Clearwater lakes are pre-
sented separately because of the potential chela-
tion by organic complexes of Ca+2 and Mg+2 in
darkwater lakes and the contribution of organic
acids. Measured values of [Ca+2 + Mg+2] in dark-
water lakes can be greater than ionic values due to
this complexation.
In all of the subregions in the Northeast, the inter-
cepts of ANC versus [Ca+2 + Mg+2] for Clearwater
lakes were negative, ranging from -72.2 in South-
ern New England (1D) to -46.3 in Maine (1E).
Slopes for Clearwater lakes (true color <30 PCU)
82
-------
Table 6-4.
Subregion
1A
1B
1C
1D
1E
2A
2B
2C
20
3A
3B
Regression Statistics for ANC (Dependent) versus Base Cations* (Independent) by Subregion, Eastern Lake Survey-
Phese I
n Regression Lakes Intercept Slope r2
114
82
81
44
32
32
123
86
86
47
38
38
93
67
67
71
16
16
83
48
48
114
83
83
84
38
38
58
35
35
55
37
37
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg]
ANC vs. [Ca+Mg+Na+K]
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
All
Clearwater
Clearwater
-57.0
-51.6
-56.8
-53.7
-46.4
6.3
-52.9
-58.2
3.5
-60.0
-72.2
44.5
-45.7
-46.3
-1.4
-39.8
-41.3
-74.9
-46.9
-57.9
-69.1
-40.8
-41.3
-20.8
-42.1
-35.5
-33.0
24.6
37.4
-0.4
-86.9
-30.9
-12.1
0.723
0.673
0.546
0.547
0.465
0.060
0.914
0.977
0.280
0.756
0.859
-0.001
0.967
0.987
0.345
0.864
0.944
0.906
0.793
0.971
0.878
0.760
0.775
0.411
0.995
0.892
0.668
0.938
0.846
0.713
0.597
0.191
-0.006
0.691
0.635
0.738
0.404
0.386
0.069
0.701
0.725
0.375
0.401
0.440
0.000
0.740
0.742
0.295
0.667
0.708
0.834
0.662
0.717
0.718
0.576
0.572
0.399
0.677
0.737
0.699
0.677
0.724
0.868
0.215
0.117
0.001
*Not corrected fo sea-salt contributions.
ranged from 0.465 in the Poconos/Catskills (1B) to
0.987 in Maine (1E). In a survey of lakes in New
England (primarily the area represented by Subre-
gions 1C, 10 and 1E), Haines and Akielaszek (1983)
found a relationship of ANC = -55 + [Ca* + Mg*],
where calcium and magnesium had been corrected
for sea-salt contributions.
The sum of base cations [Ca+2 + Mg+2 + Na+ + K+]
was regressed on ANC for Clearwater lakes. In the
Northeast, with the exception of the Adirondacks
(1 A), the addition of Na+ and K+ produced intercepts
that were near zero or were positive. In all areas the
slopes decreased and r2 values were much lower for
all areas except the Adirondacks (1 A). This indicates
that on a subregional basis the [Na+ + K*] is in
association with anions other than HCOa~, such as CI"
and SO/2, and most probably is derived from sea
spray or road salt contamination.
In the Upper Midwest, values of the intercept of
ANC versus [Ca+2 + Mg+2] ranged from -57.9 in
the Upper Peninsula of Michigan (2B) to -35.5 in
the Upper Great Lakes Area (2D). Only the slope for
Subregion 2C varied from the range of reference
values given by Henriksen (1980) and, with the ex-
ception of Subregion 2B, the subregions in Region
2 had less negative intercepts than the subregions
in the Northeast.
Also in contrast to the Northeast, when all major
cations were considered, the relationships re-
mained similar because of a lack of influence by
sea-salt. The intercepts were more negative in Sub-
83
-------
regions 2A and 2B, with a slight decrease in slope.
In Subregions 2C and 20, the intercepts were less
negative, also with a decrease in slope.
For the Southern Blue Ridge (3A), the intercepts of
[Ca+2 + Mg+2 + Na+ + K+] versus ANC for both
clearwater and all lakes were positive due to the
important contribution of sodium and potassium to
the sum of base cations. The slopes were 0.846 and
0.938 for clearwater and all lakes, respectively. Con-
sidering all major cations forced the regression
through the origin due to the high watershed
sources of sodium and potassium.
In Florida (3B), the relationship of ANC versus [Ca+2 +
Mg+2] showed poor fit for all lakes and also for
clearwater lakes. This is further evidence that Florida
(3B) lakes represent a unique set of chemical
conditions compared to lakes sampled in other
areas. Similar poor fit was observed by Wright
(1983). After noting the unusual chemical composi-
tion and unexplained high variance in sulfate con-
centrations in lakes from the Trail Ridge and High-
land Ridge areas, lakes in Florida were not analyzed
further (Wright 1983). If all major cations are used,
there is no relationship with ANC for lakes in Florida
(3B). This is related to the major contribution of
sodium to the total of base cations, and high equiv-
alents of potassium, chloride and sulfate.
5.2.5 Major Cations and Anions
For many lakes, bicarbonate and calcium are the
principal ions and much can be learned by evaluat-
ing the concentrations of these ions. However, a
more thorough understanding of lake chemistry
can be achieved by evaluating all major anions and
cations collectively. Comparison of major anions
and cations is a common technique for analyzing
lake chemistry data. An assumption in the use of
these comparisons is that all major anions and
cations have been measured.
5.2.5.1 Individual Example Lakes
The strength and focus of the Survey design is that
it allows populations of lakes to be described. How-
ever, it can be useful to evaluate differences in lake
chemistry among subregions by examining the
chemistry of individual lakes that are "typical" of a
population. No lake can be representative of other
lakes with respect to all chemical attributes, but the
relative concentrations of anions and cations ob-
served among individual lakes can be used to high-
light features of many similar lakes among sub-
regions.
A direct way of examining associations among an-
ions and cations is with ion bar charts. Figures 5-14
and 5-15 show the cationic and anionic composi-
tion of a lake selected from each of the eleven sub-
regions. Lakes were selected on a basis of their
84
ANC and DOC values to standardize the presenta-
tion such that each lake in the aforementioned fig-
ures has ANC and DOC concentrations approxi-
mately equal to the Q, (20th percentile) values for
the respective subregion (Section 4.7, Table 4-25).
The ionic concentrations of cations and anions in
each lake are expressed as percentages of the total
ionic concentration on the Y axis. The total ionic
concentration is shown adjacent to each lake. Ions
measured at less than 5 ^eq L~1 (Fe+3, Mn+2, AI+3,
NH4+, NOs", F~) are not included in this comparison.
Organic anions were not measured directly, but
their contribution to the ionic composition of the
lake can be inferred by examining the anion deficit
(sum of cations-sum of anions).
Figure 5-14 shows the cations and anions for Whit-
ney Lake, New York, in Subregion 1A. Cations
(Ca+2, Mg+2, Na+, and K+)5 are shown by the top
bar; anions (HC03~, SO*'2, and CI") by the bottom bar.
Calcium is the predominant cation, followed in order
of percent ionic concentration by magnesium, sodium,
potassium, and hydrogen ion (not shown on figure).
Sulfate is the major anion, with only small percent-
ages of bicarbonate (4.7%) and chloride (2.7%)
present. The sums of cations and anions show
close agreement and the total ionic concentration
for this lake in Subregion 1A is lowest among the
five lakes shown for Region 1.
As the figure demonstrates, Hiawatha Lake in Sub-
region 1B is considerably higher in total ionic con-
centration than Whitney Lake (1A). Lake Hiawatha
exhibited a 4 percent anion deficit, equal to
21 (xeq L~1, and showed a moderate percentage of
sodium (11.9%) and bicarbonate (15.9%). The total
ionic concentration and the relative percentages of
ions of Pemigewasset Lake in Subregion 1C are
very similar to Quabbin Reservoir in Subregion 1D,
which are primarily attributed to high concentra-
tions of sodium and chloride. Like Hiawatha Lake,
Peabody Pond in Subregion 1E also shows a 4.4
percent (16 jxeq L~1) anion deficit.
All four example lakes shown for Region 2 (Figure
5-15) exhibit some degree of anion deficit, ranging
from 5.4 percent (11 jieq L~1) in Brown Lake, Wis-
consin, to 9.9 percent (34 (xeq L~M in Louis Lake,
Minnesota. This is consistent with population esti-
mates showing high first quintile ((!,) values of
DOC for Region 2. Brown Lake, Wisconsin {Subre-
gion 2C) is noteworthy because of the low ionic
concentration of 205 jjieq L~1 and contrasts sharply
with the high ionic concentration in Blue Lake, Flor-
ida. Both Estes Lake (2B) and Brown Lake (2C) are
most similar in ionic composition to Whitney Lake
(1A), but differ in that the Region 2 example lakes
sCharges were assigned to cations by converting the units reported by the
laboratory (in mg L~1) to equivalents (Hillman et al. 1986).
-------
Figure 5-14. Ionic composition of selected lakes from Region 1, Eastern Lake Survey-Phase I.
Whitney Lake, NY
(1A1-012)
Hiawatha Lake, PA
(1B2-065)
Pemigewasset Lake, NH
(1C2-062)
Quabbin Reservoir, MA
(101 -004)
Peabody Pond, ME
(1E1-056)
Percent Total Ionic Concentration
0 10 20 30 40 50 60 T ,, .
Total Ionic
Concentration
(//eq L-1)
253
535
601
573
357
10 20 30 40 50
Percent Total Ionic Concentration
60
Cations (+)
H| Calcium ^
Anions (-)
|l Bicarbonate
N Magnesium |Oj Potassium
P^ Sulfate RS$ Chloride
J2J Sodium
show an anion deficit and a higher percentage of
potassium. Loon Lake (2D) differs from the other
lakes shown here in the high proportion of potas-
sium. This high potassium concentration is not rep-
resentative of lakes in Subregion 2D (Q-, = 15 u>eq
L~1K) and illustrates that "typical" lakes based on
some variables may not be typical with respect to
other variables.
Nottely Lake, Georgia, is typical of lakes in 3A in
exhibiting close agreement between the sums of
major ions and cations. This lake and many others
in the Southern Blue Ridge (3A) show a high percent-
age of sodium. Nottely Lake contains the lowest
percentage of sulfate and the highest percentage of
bicarbonate of all example lakes in Figures 5-14 and
5-15. Blue Lake in Florida (3B) (Figure 5-15) shows a
high total ionic concentration relative to the other
example lakes. This is attributed to the high concen-
trations of sodium and chloride. Unlike the other ex-
ample lakes in which the concentration of sodium
exceeded chloride, chloride in Blue Lake and Loon
Lake (2D) exceeded sodium. Because of the great
heterogeneity in chemistry among lakes in Subre-
gion 3B (Section 6.3), it is also likely that an exam-
ple lake could have been selected with ANC and
85
-------
Figure 6-15. Ionic composition of selected lakes from Region 2 and Subregions 3A and 3B, Eastern Lake Survey -Phase I.
Louis Lake, MN
(2A1-003)
Estes Lake, Ml
(2B1-031)
Brown Lake, Wl
(2C2-065)
Loon Lake, Wl
(2D2-064)
Nottely Lake, GA
(3A2-016)
Blue Lake, FL
(3B2-080)
Percent Total Ionic Concentration
0 10 20 30 40 50 60 Total Ionic
Concentration
(Aieq L-1)
342
274
205
350
340
1290
Cations (+)
Rza
W( Calcium
0 10 20 30 40 50 60
Percent Total Ionic Concentration
Magnesium
Potassium
Sodium
Anions(-)
I Bicarbonate
Sulfate
Chloride
DOC values similar to those of Blue Lake, but with
substantially lower concentrations of sodium and
chloride. Blue Lake, Florida, also differs from other
example lakes in the low concentration of bicarbon-
ate ion and low pH (5.35).
5.2.5.2 Order of Major Ions
The relative importance of major cations and
anions among the Subregions is summarized in
Table 5-5. The order of the ions is presented based
on population estimates for the Q1 (20th percentile)
value and the median concentration. Previous sec-
tions have described patterns of SCV2 and ANC;
this section focuses on relationships of other major
ions.
Other studies of the chemical composition of lakes
and rivers have observed the following order of
major cation concentrations: Ca + 2>Mg + 2
>Na+>K+ (Livingstone 1963; Hutchinson 1957;
Rodhe 1949; Gorham 1955). Six of the eleven ELS-I
subregions showed a rank order of cations consis-
tent with the above pattern based on the Q1 concen-
trations and only five subregions showed this order
based on the concentrations at median values.
86
-------
Table 5-5.
Subregion
Order of Major Cations and Anions Based on Population Estimates of Concentrations at the 20th Percentile (Q-|) and
Median Values, Eastern Lake Survey-Phase I
Cations
Anions
M
M
1A
1B
1C
1D
1E
2A
2B
2C
2D
3A
3B
Ca > Mg = Na > K
Ca > Mg > Na > K
Ca > Na > Mg > K
Na > Ca > Mg > K
Ca > Na = Mg > K
Ca > Mg > Na > K
Ca > Mg > Na - K
Ca > Mg > K - Na
Ca > Mg > Na = K
Ca = Na > Mg > K
Na > Mg - Ca > K
Ca > Mg > Na > K
Ca > Na =• Mg > K
Ca > Na > Mg > K
Na > Ca > Mg > K
Ca > Na = Mg > K
Ca > Mg > Na > K
Ca > Mg > Na > K
Ca > Mg > Na > K
Ca > Mg > Na > K
Ca = Na > Mg > K
Ca > Na > Mg > K
SO4>CI = HCO3
SO4>HC03>CI
SO4>HCO3>CI
CI>S04>HC03
HCO3 = SO4>CI
HC03 > SO4 > Cl
SO4>HCO3>CI
SO4>HCO3 = CI
HC03 > S04 > Cl
HC03 > Cl = SO4
CI>SO4>HC03
S04>HCO3>CI
HCO3 > S04 > Cl
HCO3 = S04 :> Cl
CI>S04=HC03
HC03>SO4>CI
HC03 > SO4 > Cl
HCO3 > S04 > Cl
HC03>S04>CI
HCO3>S04>CI
HC03 > Cl > S04
CI>S04>HC03
The rank order of anions based on median values
for seven of the subregions is HCOa" > S(>4~2 > Cl".
Only in Subregion 1A is sulfate the dominant anion
at median values. Chloride is the major anion in
Subregions 10 and 3B, consistent with the high
sodium values in these subregions. Sulfate is the
least abundant anion in Subregion 3A. The devia-
tion of major cations from the expected order is
attributed largely to the relatively high concentration
of sodium in Subregions 1D and 3B, and to a lesser
degree in Subregions 1B, 1C, 1E and 3A. The
concentration of sodium in lakes is of interest because
of the potential acidifying influence of neutral salts
(Rosenqvist 1978, Krug et al. 1984). For example,
perched coastal lakes in Australia receiving large
contributions of sea-salt have been shown to be
acidic (Bayly 1964). However, Figure 4-11 (Section
4.5.1.2) reveals that, with the exception of the Cape
Cod area, no concentrations of acidic lakes occur
near the coast.
The order of major anions represented on Table 5-5
may be somewhat misleading because not all
anions are measured in the analytical process,
which is apparent from the large anion deficits in
some lakes. By including anion deficit as an unde-
fined collection of anions (presumably composed
primarily of humic and fulvic acids), a different pat-
tern emerges (Table 5-6). At the 20th percentile of
anion concentrations, the concentration of un-
measured anions becomes second in importance in
Subregions 2A, 2C and 2D. In Subregions 2A and
20, this is attributed to the high concentration of the
unmeasured anions, whereas in Subregion 2C, this
results from the overall low ionic concentrations.
Over 20 percent of the lakes in Subregions 1A, 1B,
1D and 3A have no unmeasured anions and no sub-
regions in Region 1 show appreciable influence of
unmeasured anions at Q1 values. At median con-
centrations, all subregions except 10 and 3A show
the presence of unmeasured anions and illustrate
Table 5-6. Order of Major Anions by Subregion Based on the
20th Percentile (Qi) and Median Concentrations
Including A~ as Unmeasured Anions,* Eastern
Lake Survey-Phase I.
Subregion Q1 Median
1A
1B
1C
10
1E
2A
2B
2C
20
3A
3B
SO4>CI = HC03
S04 > HC03 > Cl
S04 > HCO3 > Cl > A-
CI>S04>HCO3
HCO3 =- SO4 > A- = Cl
HCO3 > A- = SO4 > Cl
SO4>HC03 = A->CI
SO4 > A- = HCO3 =» Cl
HCO3 > A- = S04 > Cl
HCO3 > Cl - SO4
Cl > S04 > A- = HCO3
SO4 > HCO3 > A- > Cl
HCO3 > SO4 > Cl > A-
HC03 =* SO4 > Cl > A-
Cl > SO4 =• HCO3
HCO3 > SO4 > A- = Cl
HCO3 > A- > S04 > Cl
HC03 > SO4 > A- > Cl
HC03 > SO4 > A- > Cl
HC03 > A- > S04 > Cl
HCO3 > Cl > SO4
Cl > A- = SO4 > HC03
*A~ not shown if equal to zero
the importance of considering unmeasured anions
when evaluating lake chemistry for some popula-
tions of lakes.
The moderate to high sodium (and chloride) con-
centrations in Region 1 can be attributed, in part, to
contribution from sea-salt. Figure 5-16 shows the
concentration of sodium in lakes of Region 1 as a
function of distance from the coast. Distance from the
coast is not a totally satisfactory measure of potential
contribution from marine sources because other
factors such as prevailing winds and topography
affect the transport of sea spray. The sodium concen-
tration decreases inland from the coast, approaching
minimum values for Region 1 at approximately 60 km
from the coast for sites not affected by road salt.
At least some of the lakes with high concentrations
of sodium in these lakes can be attributed to road
salt used as a deicing agent. Six states within Re-
gion 1, where approximately 1.8 million tons of
sodium chloride were applied during the winter of
87
-------
Figure 5-16. Lake water sodium concentration (jueq L~1) versus lake distance (km) from coast for Region 1, Eastern Lake
Survey-Phase I.
600-
- 400-
e
3
TJ
O
w
200-
I •
• •
•1 • V.
•
• . • V
• . I ». •.•••••
••• *
i
20
I T
40 60
Distance from Coast (km)
80
I
100
i
120
1966-67 (Field et al. 1973), have reported water pol-
lution as a result of deicing salt application (Hanes
et al. 1970). The effect of road salt and other
anthropogenic factors can confound the relationship
shown in Figure 5-16. This can be minimized by
focusing on the lowest concentrations of sodium as a
function of distance from the coast. In this case, the
minimum concentrations of sodium show a sharp
decline near 20 km and approach a concentration of
about 30 yt/eq L~1 at approximately 40 km from the
coast. This agrees with Ogden (1982) who observed
that about 80 percent of the marine aerosol in Nova
Scotia is deposited within 20 km of the coast. Jackson
(1905), Mairs (1967) and Haines and Akielaszek
(1983) observed a similar pattern using chloride, but
88
they observed that the influence of marine sources
did not reach background concentrations until 75 to
100 km from the coast.
The relationship between sodium and distance
from the coast for lakes in Subregion 3B (Florida) is
presented in Figure 5-17. Only two of the sample
lakes within 20 km of the coast in Florida (including
the southern portion of Georgia) have sodium val-
ues exceeding 300 jxeq L"1 and in contrast to Re-
gion 1, the lakes in Florida exhibit a positive rela-
tionship between sodium and distance from the
coast (slope = 5.0). This occurs despite the obser-
vation that the concentration of sodium in the pre-
cipitation appears to be low for the area containing
-------
Figure 5-17. Lake water sodium concentration (/jeq L 1) versus lake distance (km) from coast for Subregion 3B, Eastern Lake
Survey-Phase I.
1
E
3
700-
600-H
500-
400-
300-
200-
100-
20
I I I
40 60 80
Distance from Coast (km)
100
120
most of the study lakes (Environmental Science and
Engineering, Inc. 1983). This suggests that many of
these lakes have a substantial contribution of
sodium from the watersheds.
Further support for the idea that much of the
sodium is not of atmospheric marine origin is
shown by comparing sodium and chloride concen-
trations in Subregion 1D (Southern New England)
with those in Subregion 3B (Florida, Figure 5-18). A
line showing the expected fit of 0.86, based on the
molar ratio of sodium to chloride in the ocean, is
superimposed over the data. Lakes in Subregion 1D
show good agreement with the expected ratio of
sodium:chloride (slope = 0.827, r2 = 0.959),
whereas lakes in Subregion 3B show a slope of
0.642 and greater variability in the ratio (r2 = 0.796).
Lakes affected by road salt in Subregion 1D cannot
be distinguished from sea spray sources on the
basis of this relationship.
Table 5-7 shows the regression statistics for
sodium versus chloride for all subregions. The de-
gree of association between the variables is
strongest in Region 1 where the regression gener-
ally explains over 94 percent of the variance, and
the slope is generally close to the expected ratio of
sea-salts. Lakes in Region 2 show greater scatter in
the relationship, although this appears to be related
to the low concentrations of sodium and chloride in
this region (Section 4.8., Table 4-30). Subregion 3A
shows a far greater ratio of sodium :chloride than
89
-------
Figure 5-18. Relation of sodium (yueq L~1) to chloride (/ueq L'1) for lakes in Subregions ID (A) and 38 (8), Eastern Lake
Survey-Phase I. (The line represents the expected ratio (a slope of 0.86) of sodium to chloride in the ocean.)
1000
800-
600-
1
E
3
1 400
w
200-
200 400 600 800 1000
Chloride (yueq L~1)
Chloride (/ueq L"1)
90
-------
Table 6-7.
Subregion
Regression Statistics for Sodium (Dependent)
versus Chloride (Independent) by Region for
Concentrations from 0 to 1000 Aieq L~1, Eastern
Lake Survey-Phase I
Intercept
Slope*
1A
18
1C
1D
1E
2A
2B
2C
2D
3A
3B
200
152
204
112
183
158
155
187
141
110
153
25
10
34
32
30
33
15
17
24
10
45
0.811
0.838
0.876
0.827
0.938
0.710
0.825
0.705
0.713
1.600
0.642
0.898
0.954
0.950
0.959
0.970
0.815
0.722
0.837
0.723
0.780
0.796
•Expected sea-salt ratio = 0.86 (Holland 1978)
observed in other subregions. The most plausible
explanation for the high sodium:chloride ratio in
3A is weathering of the shale common to this area.
Deviations from the expected sea-salt ratio of 0.86
in precipitation and lakes have been attributed to
anthropogenic sources (e.g., road salt, sewage,
agriculture), exchange of marine sodium for diva-
lent cations (Thompson 1982), exchange of sodium
with H+ (Rosenqvist 1978), contributions of excess
sodium or chloride from the soil (Junge and Werby
1958, Lebowitz and de Pena 1985), and loss of chlo-
ride particles by volatilization of chloride as HCI
(Eriksson 1960).
In view of the positive relationship between sodium
and distance from the coast in Subregion 36, the
most likely dominant source of the high sodium
(and other major ions) in many Florida lakes is con-
tribution from shallow groundwater aquifers.
Heath and Conover (1981) report that the median
conductivity of public water supplies derived from
shallow sand aquifers in Florida is 520 |xS cm"1
(range 220 to 988). However, a more complete eval-,
uation of the chemistry of lakes, groundwater, and
deposition is required before this issue can be re-
solved.
Although the concentration of calcium exceeds
magnesium in all subregions, except Subregion 3B,
the relationship of calcium to magnesium is highly
variable among and within subregions. Figure 5-19
shows plots of calcium versus magnesium for
Northcentral Wisconsin (2C) and Florida (3B) and
illustrates extreme cases of this relationship among
'the subregions. Lakes in Subregion 2C are located in
a similar geologic setting almost entirely within the
Wisconsin Valley glacial lobe. This relative homo-
geneity of surficial geological material probably
contributes to the high degree of correlation (r2 =
0.891). In contrast, the lakes in Florida (3B) are
located over a variety of different soil types and
surficial and bedrock geology, and have groundwater
contributions (Heath and Conover 1981). The regres-
sion results for calcium and magnesium are shown
for all subregions in Table 5-8. Only Subregion 3B
exhibits a slope of 2. The degree of association between
calcium and magnesium is quite high in Subregions
2B, 2C and 2D. The greater degree of scatter observed
in Subregion 2A is most likely attributable to the
variety of bedrock source material.
Table 5-8.
Subregion
Regression Statistics for Calcium (Dependent)
versus Magnesium (Independent) by Region for
Concentrations from 0 to 1000 peq L~1, Eastern
Lake Survey-Phase I
Intercept
Slope
1A
1B
1C
1D
1E
2A
2B
2C
2D
3A
3B
196
150
201
124
181
158
145
187
124
110
137
33
72
59
14
23
17
16
-4
-14
5
85
2.364
1.706
1.946
1.479
2.373
1.514
1.907
1.756
1.685
1.650
0.579
0.756
0.554
0.539
0.594
0.451
0.471
0.836
0.891
0.948
0.641
0.442
5.2.5.3 Relationships Among Major Ions
The previous descriptions of lake chemistry among
subregions have been presented largely on the
basis of absolute concentrations. Another useful
approach to evaluate populations of lakes, particu-
larly for anions and cations, is to compare ratios of
ions. Figures 5-20, 5-21 and 5-22 present ratios of
major anions to major cations as trilinear plots
(ternary diagrams) (Hem 1970). Like the cation/
anion bar charts (Figures 5-14 and 5-15), these plots
include only the major ions and exclude ions that
may be important in a small number of these lakes.
These ions with typically low concentrations are
discussed further in Section 5.2.6. The axes on the
plots represent the percentage concentration of the
ion or pair of ions, where values range from 0 to 100
percent, increasing in the direction of the arrows. A
point representing the percentage ion composition
for a lake is plotted once on each triangle for the
cations and anions and is plotted a third time on the
parallelogram for the pairs of cations and anions.
Thus a lake with 100 percent Ca(HCO3)2 would ap-
pear in the extreme lower left portion of each trian-
gle and would appear in the left corner of the paral-
lelogram. A lake acidified by su If uric acid would be
expected to appear near the apex of the parallelo-
91
-------
Figure 5-19. Relationship of calcium (/ueq L~1) to magnesium (/ueq L'1) for lakes in Subregions 2C (A) and 3B (B). Eastern Lake
Survey-Phase!.
1000
cr
1
800-
600
400-
200-
200
i r
400 600 800 1000
Magnesium (/^eq L"1)
1000
cr
1
E
800-
600-
400-
200-
0 200 400 600 800 1000
Magnesium (/jeq L"1)
92
-------
Figure 5-20. Trilinear plots of major anions and cations in Subregions 1 A, IB, 1C and 1D, Eastern Lake Survey-Phase I.
Cl
gram. Although these trilinear plots suffer from two
deficiencies (i.e., they neither show differences in
concentrations, nor do they include other ions of
possible importance), they are useful for displaying
patterns among the major ions.
There is considerable scatter present in the rela-
tionships of ions among and within regions. The
ratios of anion concentrations generally exhibit a
far greater degree of variability than observed for
the ratios of cation concentrations. This is consis-
tent with the observation discussed in Section
5.2.5.2 (Table 5-5) showing that the order of domi-
nant anions changed with increasing concentra-
tions. The ratios of cation concentrations show rel-
atively little variability in Subregions 1A, 1B, all of
Region 2, and Subregion 3A. The contribution of
road salt, sea-salt, and groundwater sources of
sodium is evident in Subregions 1C, 1D, 1E and 3B.
The plots for anions exhibit several distinct pat-
terns. Subregions 1A, 2A, 2B, 2C and 2D could be
classified as relatively high sulfate, low chloride
waters. The bicarbonate concentrations are in
some cases very low in Subregions 1A, 2B and 2C
and the lakes in these three Subregions show a high
degree of similarity of anion ratios. Subregions 1D
and 3B are high chloride areas with a moderate
degree of variability in anion concentration ratios.
Subregion 3A is the only subregion in which the
93
-------
Figure 5-21 Trilinear plots of major anions and cations in Swbregions 1E, 2A, 2B and 2C, Eastern Lake Survey -Phas
lakes are consistently bicarbonate systems with
low concentrations of organic acids. Subregions
1B, 1C, and 1E are distinct from the other subre-
gions because of the lack of any distinct pattern in
anion composition.
5.2.5 Dissolved Organic Carbon
5.2.6.1 Color and DOC
Water passing through plant materials leaches
complex organic compounds that are resistant to
decomposition. These organic compounds often
contain quinoid functional groups that impart a
brown color to water (Gjessing 1976). Because of
this property, color often has been used as a surro-
gate for DOC.
Regression equations were computed to evaluate the
relationship between DOC and true color for lakes in
each subregion (Table 5-9). The intercepts ranged
from 1.66 to 2.59 mg L~1 DOC for subregions in the
Northeast, whereas the intercepts for the Upper
Midwest ranged from 2.20 to 5.69 mg L'1 DOC. Only
the Southern Blue Ridge (3A) had an intercept not
significantly different from zero. Three subregions
with large proportions of low ANC lakes 1 A, 2C and
3B, had similar intercepts of approximately 2 mg L~1
DOC. Intercepts greater than zero indicate that some
94
-------
Figure 5-22. Trilinear plots of major anions and cations in Subregions 2D, 3A and 3B, Eastern Lake Survey -Phase I.
of the organic matter does not impart color to the
water. This is consistent with the observations of
Lamar and Goerlitz (1966) who isolated colorless
carboxylic acids from stream water samples.
The slopes of the regressions between DOC and
color ranged from 0.049 in the Southern Blue Ridge
(3A) to 0.124 in Florida (3B), although most subre-
gions showed slopes from 0.07 to 0.10. The low r2
value for Subregion 1C is, in part, attributed to a
possible analytical error for one DOC measure-
ment; excluding this value from the analysis yields
an intercept of 2.49, a slope of 0.083, and an r2 of
0.564. A typical relationship between DOC and
color is shown in Figure 5-23 for lakes in Maine (1E).
Considerable scatter is evident such that for a given
color value, the DOC concentration may vary by
several milligrams per liter. Such variations in ob-
served DOC concentrations are expected because
of the effect increasing pH has on increasing the
color of surface waters (Black and Christman 1963).
The lakes in Northeastern Minnesota (2A) exhibit
an even greater degree of scatter between DOC and
color, particularly at high color (Figure 5-24). The
lowest DOC value in Subregion 2A is a suspected
analytical error; removal of this observation from
the regression results in an intercept of 5.21, a slope
of 0.075, and an r2 of 0.587. However, this does not
explain the relatively poor agreement between
DOC and color in Northeastern Minnesota (2A).
95
-------
Figure 5-23. Dissolved organic carbon (0-20 mg L~1) versus color (0-200 PCU) for Subregion 1 E, Eastern Lake Survey-Phase I.
20
16 —
12
o>
u
o
Q
8 —
4-
. . ::\
t *
s •
;» . •
i.i-:
:s!«: •.
I
40
I
80
120
160
200
Color (PCU)
There is a variety of factors that could contribute to
poor agreement between DOC and color (e.g., ef-
fect of pH, different sources of DOC, different resi-
dence times in the lakes), but no single factor pro-
vides a satisfactory explanation in this relatively
homogeneous area (Section 6.2).
The extremely small slope for lakes in Subregion
3A (Figure 5-25) can be attributed to the low con-
centrations of organic acids in these waters (as indi-
cated by the lack of any appreciable anion deficit).
The most probable source of color in these lakes is
from light scattering caused by inorganic colloids.
The slope for Florida lakes (3B) is significantly greater
than for other subregions, indicating that for a given
DOC the expected color is low compared to other
subregions. Florida lakes are exposed to more inten-
96
sive sunlight and have a higher mean annual
temperature than lakes in Regions 1 and 2. Both
ultraviolet radiation (ll'in and Orlov 1973) and higher
temperatures (Gjessing 1976) contribute to a reduc-
tion in color. Gjessing also points out that color
reduction of humic acids can occur with increasing
salt concentrations. However, the small slope in
Southern New England (1D), a subregion with equally
high sodium chloride concentrations, suggests that
the effect of salt on color reduction on a regional basis
may be small. Finally, vegetation differences between
Florida and other subregions could contribute to a
different molecular composition of humic and fulvic
acids.
Although there is a general relationship between
, DOC and color, the relationship is highly variable
-------
Figure 5-24. Dissolved organic carbon (0-20 mg L 1) versus color (0-200 PCU) for Subregion 2A, Eastern Lake Survey -Phase I.
20
16 -
12 —
o>
O
o
D
8 -
4 -
40
80
I
120
I
160
200
Color (PCU)
among subregions. A number of reasons may ex-
plain the difference of the relationship in Florida
compared to those in the Northeast and the Upper
Midwest, but no simple explanation is evident to
explain the variability within Regions 1 and 2.
5.2.6.2 Anion Deficit
The principle of electroneutrality states that the
equivalent sum of cations must equal the equiva-
lent sum of anions if all contributing ions are mea-
sured. In addition to serving as an analytical tool for
quality assurance checks, evaluation of ion bal-
ances provides additional insight into regional lake
chemistry. Many of the lakes sampled had anion
deficits; i.e., the sum of measured anions was less
than the sum of measured cations. Assuming no
analytical error, the deficit reflects an inability to mea-
sure organic anions directly. It can be inferred that
the anion deficit is caused by unmeasured organic
anions if a relationship between anion deficit and
DOC can be established. Table 5-10 shows the pop-
ulation estimates, the quintiles Q, and Q4, and me-
dian for anion deficit by subregion. Subregion 3B
contains the highest estimates of anion deficit at
the median and Q4, whereas Subregion 2A contains
the highest estimate at the 20th percentile (Q,).
Lakes in the Upper Midwest (with the possible ex-
ception of Subregion 2C), contain substantially
higher estimates of anion deficit than lakes in the
Northeast. Subregions 1A, 1B, 1D and 3A have no
anion deficit in over 20 percent of the lakes. Caution
should be used in interpreting the population esti-
mates of anion deficit because subregions with
high ionic concentrations (such as Florida) can ex-
97
-------
Figure 5-25. Dissolved organic carbon (0-20 mg L 1) versuscolor (0-200 PCU) for Subregion 3A, Eastern Lake Survey -Phase I.
20
16 -
12-
o>
(J
O
Q
8 -
4 -
.•!'•• •
Ii.! !:!'« : • •
I
40
I
80
120
I
160
200
Color (PCU)
hibit a deceptively high anion deficit related to
greater absolute analytical error at high ionic con-
centrations.
There is a strong linear relationship between the
sums of anions and cations among all regions and
within most subregions. Summary statistics for re-
gressions between the sum of major anions and
cations are shown in Table 5-11 for darkwater
(color > 30 PCU) and clearwater lakes (color
<30 PCU). Complete agreement between mea-
sured anions and cations would result in an inter-
cept of zero and a slope of one. Most lakes show an
intercept and slope less than one, indicating a sys-
tematic underestimation of the anions. This pattern
is even more apparent in the darkwater lakes. The
subregions in the Upper Midwest generally exhibit
the greatest departure from the 1:1 relationship.
Furthermore, the slopes for lakes in the Upper Mid-
west and Florida show a substantial decrease in
slope between the clearwater and darkwater lakes
that indicates an increasing concentration of un-
measured anions at higher ionic concentrations.
This is in contrast to the pattern for lakes in the
Northeast that shows an additive effect to the anion
deficit in the darkwater lakes (i.e., the slopes be-
tween clearwater and darkwater lakes are similar,
with only the intercepts showing a substantial dif-
ference).
Intercepts greater than one indicate a cation deficit
caused by not measuring all cations (assuming no
analytical error). In the Northeast, only Subregion
1A shows a positive intercept for both clearwater
98
-------
Table 5-9.
Subregion
Regression Statistics for DOC (0-20 mg L~1, De-
pendent) versus Color (0-200 PCU. Independent)
toV Subregion, Eastern Lake Survey- PnM8 ,
r2
Intercept
Slope
1A
1B
1C+
1D
IE
2A*
2B
2C
20
3A
3B
203
154
213
127
181
151
152
185
141
112
141
1.90
1.66
2.59
1.77
2.46
5.69
3.08
2.20
3.75
0.25
1.99
0.081
0.093
0.072
0.077
0.096
0.064
0.089
0.094
0.085
0.049
0.124
0.770
0.615
0.451
0.753
0.732
0.372
0.541
0.785
0.663
0.469
0.772
+ Removal of one observation results in the following: intercept
2.49, slope 0.083, and r2 0.564.
•Removal of one observation results in the following: intercept
5.21, slope 0.075, and r2 0.587.
Table 5-10. Population Estimates of Q-\ (20th Percentile),
Median and O* (80th Percentile) Anion Deficit*
(ueq L~1) by Subregion, Eastern Lake Survey-
Phase I
Subregion Q, M Q4
1A
IB
1C
1D
1E
2A
2B
2C
2D
3A
3B
0
0
3
0
13
48
20
12
29
0
12
24
13
26
0
42
78
55
34
81
0
98
49
45
60
35
85
118
150
73
148
16
187
* Anion deficit is expressed as the positive difference of 'S,
cations-2 anions; where 2 anions a 2 cations, anion
deficit = 0; and where the cations are Ca+2, Mg+2, Na+, K+,
NH«+, H+, and the anions are HC03~, CDs'2, OH", S04~2, Cl~, F",
N03-.
and darkwater lakes. A plot of the sum of anions
versus sum of cations for lakes in the Adirondacks
(1 A; Figure 5-26) shows a group of dilute lakes with
a cation deficit. Including Al+3, Fe+3 and Mn+2 in the
sum of cations resolves this anomaly and results in
the following regression statistics for lakes in Sub-
region 1 A: intercept -8.3, slope 0.890, and r2 0.966.
Most subregions show an intercept near zero and a
slope slightly less than one for clearwater lakes.
The regressions for the darkwater lakes show a
substantially lower intercept for all subregions ex-
cept for Subregion 3A (Southern Blue Ridge). The
close agreement between measured sums of
anions and cations in Subregion 3A for both clear-
water and darkwater lakes is expected considering
the low DOC concentrations in these lakes (Figure
5-27). The lakes in Northeastern Minnesota (2A) are
in strong contrast to those in the Southern Blue
Ridge (3A) with respect to the relationship between
anions and cations (Figure 5-28). It is evident that
lakes through the range of ion concentrations ex-
hibit a substantial anion deficit in Subregion 2A.
Subregion 1D (Southern New England) is the only
area with a slope greater than one for both clear-
water and darkwater lakes. This subregion con-
tained the lowest estimated concentrations of un-
measured anions among lakes in Region 1.
In order to evaluate further the nature of the un-
measured anions among subregions, the anion
deficit was regressed against DOC. Table 5-12
shows the regression results computed including
the anions and cations used in previous presenta-
tions and also computed with the addition of Al+3
and Fe+3 to the sum of cations. There is some un-
certainty regarding the appropriate charge per ion
for metals in lakewaters (Driscoll et al. 1983). The
charge values of +3 were selected for this analysis
to provide a basis for comparison to the results of
Oliver et al. (1983). The addition of total aluminum
Table 5-11.
Subregion
Regression Statistics for the Sum of Anions (Dependent) versus the Sum of Cations* (Independent) by Subregion,
Eastern Lake Survey-Phase I
Clearwater Lakes Darkwater Lakes
Intercept
Slope
Intercept
Slope
1A
1B
1C
1D
1E
127
103
125
66
101
23.9
4.2
-5.8
-5.8
6.5
0.866
0.987
0.960
1.042
0.901
0.981
0.974
0.980
0.978
0.971
60
38
60
28
76
7.4
-28.8
-27.6
-42.0
-54.1
0.843
0.987
0.941
1.043
0.946
0.967
0.975
0.971
0.978
0.933
2A
2B
2C
2D
42
71
117
45
-32.6
-0.3
-7.3
-10.6
0.958
0.933
0.944
0.907
0.988
0.960
0.992
0.991
109
59
60
65
-55.2
-31.5
-35.2
-41.1
0.873
0.837
0.876
0.900
0.944
0.897
0.963
0.945
3A
3B
59
49
4.0
6.5
0.980
0.914
0.963
0.981
46
42
0.5
-59.6
1.018
0.865
0.981
0.917
•Anions included are HC03~, OH", C03"2, S04"2, CI", NO3", F;
Cations included are Ca*2, Mg*2, Na*, K*, H*, NH4+.
99
-------
Figure 5-26. Relationship of the sum of anions {/^eq L'1) to the sum of cations (//eq L"1) for lakes in Subregion 1 A, Eastern Lake
Survey-Phase I.
1000
800-
600-
1
400 —
200-
200
800
1000
Sum of Cations (fieq L'
and iron to the sum of cations improves the agree-
ment between anion deficit and DOC in all subre-
gions except Florida (3B). Despite this improve-
ment, the DOC explains less than 50 percent of the
variance for anion deficit in seven of the eleven
subregions. Better agreement between anion
deficit and DOC generally occurs in subregions with
moderate to high concentrations of DOC.
The slope of the relationship between anion deficit
and DOC has been used by others to estimate the
carboxylic acid content of the organic material
(Eshleman and Hemond 1985; Oliver et al. 1983).
The reported conversions for DOC to carboxylic acid
in surface waters range from 4.78 //eq/mg C in bog
waters (Gorham et al. 1985) to 22 //eq/mg C in the
Satilla River, Georgia (Beck et al. 1974). Oliver et al.
700
(1983) reported an average of 10.5 //eq/mg C based
on a variety of waters in North America. The
conversion estimated here (using Al*3 and Fe*3)
ranges from 6.3 //eq/mg C in Subregion 3A to 11.8
//eq/mg C in Subregion 1A. Use of 10//eq/mg C as
proposed by Oliver et al. (1983) for all lakes may
substantially overestimate (or possibly underestimate
for some areas) the organic anion concentrations
depending on the Subregion.
5.3 Hydrology
5.3.1 Hydrologic Lake Type
The physical features of a watershed and lake con-
trol how water is collected and routed to a lake, and
-------
Figure 5-27. Relationship of the sum of anions (//eq L"1) to the sum of cations f/jeq L"1) for lakes in Subregion 3A, Eastern Lake
Survey-Phase I.
1000
800-
600-
CT
3
01
C
o
'c
E
^
w
200-
400 -I
1000
Sum of Cations (/^eq L"1)
how rapidly water is replaced in a lake. Therefore,
watershed hydrology can play a major role in deter-
mining the chemical composition of a lake. Lake
chemistry represents the integration of the chem-
istry of deposition and the influence of factors such
as watershed vegetation, soils, geology and in-lake
processes.
The influence of hydrology on the chemistry of
lakes in the ELS-i was evaluated through two ap-
proaches. First, lake types were classified based on
the presence of permanent surface water inlets and
outlets identified on 1:24,000-, 1:25,000-, and
1:62,500-scale USGS topographic maps. The lakes
were classified as follows: seepage (no inlets, no
outlet); drainage (outlet); closed (inlets, no outlet);
and reservoir (outlet control structure present). The
objective of this initial classification was to distin-
guish those lakes that receive substantial chemical
input from the watersheds (e.g., drainage, closed
and reservoir lakes) from those that receive mini-
mal contributions from the watersheds (e.g., seep-
age lakes).
The hydrologic interactions between lake and
watershed are subject to considerable temporal
and spatial variability (Winter 1977). The nature of
the interaction is affected by many watershed char-
acteristics that are likely best represented by a con-
tinuum rather than discrete classifications. Al-
though the hydrologic classifications inadequately
represent the complex nature of hydrologic contri-
101
-------
Figure 5-28.
Relationship of the sum of anions (fjeq L 1) to the sum of cations (jueq L 1) for lakes in Subregion 2A, Eastern Lake
Survey-Phase I.
1000
800-
600-
CT
1
U
c
o
o
E
CO
400-
200-
200
I
I
400 600
Sum of Cations (//eq L"1)
800
1000
butions to many lakes, they provide a useful frame-
work for exploring relationships among variables.
Population estimates for the number of lakes pre-
sented in Table 5-13 revealed marked contrasts in
the abundance of lakes by hydrologic types among
subregions. Region 1 showed a high percentage of
drainage lakes and Subregions 2C, 2D, and 3B
showed a high percentage of seepage lakes. In the
Upper Midwest, Subregions 2A and 2B contained a
majority of drainage lakes (74 and 51%, respec-
tively). In the Southern Blue Ridge (3A), an esti-
mated 90 percent of the lakes were reservoirs, a
feature in strong contrast to other subregions. Only
Subregions 1B (Poconos/Catskills) and 1D (South-
ern New England) contained significant percent-
ages of reservoirs (42 and 29%, respectively). The
drainage lakes were the most common type in the
Northeast, whereas seepage lakes were common in
the Upper Midwest and Florida. Closed lakes (those
containing an inlet, but no permanent surface water
outlet) were relatively uncommon in all regions.
Population estimates of the median values of six
primary chemical variables are also shown in Table
5-13. Estimates of median extractable aluminum
are shown, but these should be interpreted with
caution because the median values were near the
system decision limit of 8 ixg L~1. When the sample
size for a lake type with in a region was less than 10,
estimates of the lake chemistry are not provided.
Seepage lakes exhibited lower median ANC values
than drainage lakes in all subregions; typically
seepage lakes also had lower Ca+2 and SO4~2 con-
702
-------
Table 6-12.
Subregion
Regression Statistics for Anion Deficit (0-200 jwq L~\ Dependent) versus DOC (0-200 mg L~
Computed without Metals (Al+3, Fe*3) and with Metals, Eastern Lake Survey-Phase I
Independent)
Anion Deficit8
Excluding Fe+3 and Al+3
Anion Deficit13
Including Fe+3 and Al+3
Intercept
Slope
Intercept
•Anion deficit computed using: Ca*a, Mg*2, Na+, K+, H+, NH«+; HCOj', OH', COs'2, Cf, NOs', F'
"Cations included are: Ca*2, Mg+z, Na*, K*, H+, NH4+, Al+3, Fe*3; (anions same as in a).
Slope
1A
1B
1C
1D
1E
2A
2B
2C
2D
3A
3B
196
154
208
120
181
148
141
184
131
110
126
-13.13
-9.31
1.85 ,
-10.18
-7.95
18.90
-4.81
-11.10
-14.04
4.87
6.12
9.45
7.06
5.94
5.61
9.42
6.32
7.91
9.36
9.26
2.30
8.14
0.330
0.315
0.206
0.328
0.620
0.390
0.472
0.651
0.520
0.037
0.433
196
154
207
120
180
148
140
179
129
110
120
-10.39
-8.99
4.06
-15.06
-7.25
18.67
-5.93
-12.72
-15.63
6.34
10.90
11.82
8.41
7.59
8.04
11.04
7.90
9.58
10.91
10.50
6.31
7.89
0.469
0.343
0.303
0.483
0.681
0.426
0.553
0.651
0.579
0.130
0.383
centrations. Median DOC values were also lower in
seepage lakes compared to drainage lakes among
all subregions except 1C and 2A.
The differences in median ANC between drainage
and seepage lakes within subregions ranged from
only 3 jj-eq L~1 in Subregion 1A to 910 u-eq L~1 in
Subregion 2D. These differences in ANC were gen-
erally much greater between drainage and seepage
lakes in Region 2 and Subregion 3B and were rela-
tively small for those lakes in Region 1. This sug-
gests either a greater volume of groundwater in-
flow or that the chemical composition of
groundwater in Region 2 and Subregion 3B con-
tributes greater ANC to the drainage lakes com-
pared to drainage lakes in Region 1. The results for
median ANC in the Adirondack(1 A) lakes are notable,
not only because there is little difference between
seepage and drainage lakes, but because the ANC
of drainage lakes in Subregion 1A is also the lowest
among drainage lakes across subregions.
Reservoirs in all subregions had among the highest
median ANC and Ca+2. This was expected in reser-
voirs because of the large size of the watersheds
relative to the lake area. ANC values for closed lakes
were most similar to those for reservoirs in Region
1, but were most similar to those for seepage lakes
in Region 2 and Subregions 3A and 3B.
The relative importance of hydrologic lake types
among the regions is contrasted in Figure 5-29. The
estimated number of lakes in each region is indi-
cated by the total height of the vertical bars. The
number of lakes with ANC >200 u,eq L~1 is shown
in white. The hydrologic lake types are represented
for lakes with ANC <200 jieq L~1. Of 7096 lakes (>4
ha and <2000 ha; Table 4-12) in Region 1, the ma-
jority (4258; Table 4-15) have ANC =£200 u,eq L~1. Of
the lakes in Region 1 with ANC <200 |xeq L~\ the
vast majority are drainage lakes. The relative pro-
portions of low ANC drainage lakes to seepage
lakes observed in Region 1 are reversed in Region
2 and Subregion 3B. This apparent regional differ-
ence in the hydrology of low ANC lakes highlights
the need to include hydrology in assessing the im-
pacts of acidic deposition.
5.3.2 Hydraulic Residence Time
Use of hydrologic lake types is a qualitative ap-
proach for evaluating the relationship between lake
chemistry and hydrology. A more quantitative ap-
proach to evaluating the relationship between hy-
drology and lake chemistry is to compute hydraulic
residence time.6 This is defined as the time required
to exchange the volume of water in a lake.
Hydraulic
Residence = -— .
Time Total Volume of Inflow per Year
Lake Volume
The hydraulic residence time of a lake can be im-
portant in response to acidic deposition for several
reasons. A lake with a short residence time would
presumably be more likely to respond to episodic
events such as snowmelt. Also, one would expect
such lakes to achieve a steady-state with respect to
a given pollutant loading relatively quickly. In con-
trast, lakes with long hydraulic residence times
likely would be slower to respond to changes in
loading assuming other watershed features were
should not be confused with retention time, which is the average
time a substance would remain in the lake.
703
-------
Table 5-13.
Subregion/
Region
1A
1B
1C
1D
1E
1
2A
2B
2C
2D
2
3A
3B
Population Estimates of Lake Numbers by Hydrologic Types with
Parameters, Eastern Lake Survey-Phase I
Estimated
Lake Type
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Seepage
Closed
Reservoir
Drainage
Closed
Reservoir
Drainage
Seepage
Closed
Number of
Lakes
991
110
89
100
799
58
622
1206
102
105
70
690
167
80
381
1386
41
56
43
5072
478
330
1216
1074
232
119
31
539
396
64
52
529
876
60
15
2080
2142
208
84
4222
3646
451
182
18
7
232
452
1380
265
PH
6.58
6.50
7.03
7.02
6.96
7.12
6.78
6.36
6.92
6.74
6.77
6.47
6.91
6.96
6.92
6.84
6.51
6.99
7.08
6.97
6.79
6.74
7.38
6.49
6.85
7.37
6.17
7.50
7.03
6.74
7.40
6.76
6.79
6.96
6.99
6.82
6.40
6.48
ANC
(|ieq L-1)
88
85
177
234
317
250
118
72
130
145
136
69
260
256
151
145
84
188
252
198
152
122
651
87
574
517
36
1224
314
94
761
120
138
181
253
190
79
68
DOC
(mg L^1)
4.3
3.4
3.7
3.4
4.0
3.6
4.5
5.4
4.4
4.1
3.9
3.0
2.7
4.5
5.2
-- Insufficient
4.4
4.0
3.9
3.8
9.0
8.5
10.7
InQi iff ifipnt
IllOUIIIlrlClll
8.3
4.8
2.6
•- Insufficient
5.8
4.1
InQi iff ipipnt
IlloUlllUldll
•- Insufficient
9.4
6.9
6.0
In^i iff ipipnt
IMOUIIIUICIIl
9.0
6.1
7.5
8.1
•- Insufficient
•- Insufficient
2.0
10.2
7.4
7.8
Associated Median Values of
Ca+2
-------
Figure 5-29. Population estimates of lake numbers with ANC <200 //eq L"1 by hydrologic type for all regions. Eastern Lake
Survey-Phase I.
9000
8000
7000
£
z 6000
ra
CO
LLJ
C
o
Q.
£
5000
4000
3000
2000
1000
I I ANO200
Reservoir
Closed
Y///////A Seepage
l::::::::::::::;:::*xl Drainage
2 3A
Region
3B
similar. Also, in-lake processes perhaps would ex-
ert a greater influence over chemistry in lakes with
long hydraulic residence times.
A large, deep lake with a relatively small watershed
would be expected to have a long (e.g., >3 yr) hy-
draulic residence time whereas a shallow lake with
a large watershed would be expected to have a
short (e.g., <0.5 yr) hydraulic residence time. Hy-
draulic residence time was approximated for the
lakes in the ELS-I using the following expression:
RT =
LA x (Zs x 0.464)
where:
[RO x (WA - LA)] + (LA x PRECIP)
RT = hydraulic residence time (yr)
LA = lake area (ha), measured from
USGS topographic maps
Zs = site depth (m), measured in the
lake
RO = runoff (m/yr), interpolated from
national runoff maps
WA = watershed area (ha), measured
from USGS topographic maps
PRECIP = precipitation (m/yr), interpolated
from National Weather Service
data.
The numerator of the equation represents lake vol-
ume which was computed by multiplying lake area
by mean depth, the latter of which was approxi-
mated by multiplying measured site depth by 0.464
(Wetzel 1983). This constant (0.464) for relating
mean depth to maximum depth varies with the
shape of a lake basin. Runoff was estimated for
each lake's catchment (WA - LA) using a linear in-
terpolation from mapped contours of mean annual
runoff (Busby 1966). Precipitation at each lake was
also interpolated from precipitation maps (Environ-
mental Data Service 1983).
There are several potential sources of error in this
estimation of hydraulic residence time (RT) in addition
to errors associated with measurements of lake
105
-------
area, watershed area, and interpolation of runoff
and precipitation. Site depth is used here as a surro-
gate for the deepest point in the lake, but not all
lakes in this Survey were sampled at maximum
depth. Failure to measure the maximum depth
would cause the hydraulic residence time to be
underestimated. Use of this approximation also as-
sumes that lake volumes are at steady-state and
that inflow and outflow are equal. Another as-
sumption is that the topographic divide is coinci-
dent with the actual surface and groundwater di-
vides. This assumption may be violated for lakes
where contributing areas are independent of topo-
graphic boundaries, a problem that can be particu-
larly severe for seepage lakes (Schnoor et al. 1985).
For this reason, RT was not computed for seepage and
closed lake types.
It was not possible to measure watershed
boundaries accurately for many Florida lakes be-
cause the slight variations in topographic relief
were undetectable on large-scale maps. In addition,
large fluctuations in lake stage cause inlets and out-
lets for many Florida lakes to be ephemeral (Heath
and Conover 1981) and result in unacceptable un-
certainty in estimates for RT. Consequently, RT could
not be computed effectively for lakes in Florida (3B).
The calculated RT for lakes in Regions 1 and 2, and
Subregion 3A is shown in Table 5-14. Lakes in the
Southern Blue Ridge (3A) have a short RT, a feature
typical of reservoirs. Lakes in the Northeast exhibit
estimated median RT values twice as high as those in
Subregion 3A, yet, except for the Q4 values of 1.03 yr
in Subregion 1B, all quintiles shown for Region 1 are
less than one year. The similarity in RT among the
lakes in Region 1 suggests that variation in lake
chemistry among these subregions may not be attrib-
uted to differences in RT. This does not exclude the
possibility that chemical differences among lakes in
Region 1 are related to hydrology. Flow paths from the
watershed to the lakes may be important (Chen et al.
1984) but are not considered in this analysis.
The RT values for lakes in Region 1 contrast sharply
with those in the Upper Midwest where Q« values for
RT exceeded one year in all subregions. Recall that RT
was not computed for seepage and closed lake types;
these lake types account for approximately one-half
of all lakes in the Upper Midwest. If it were possible to
estimate the RT values for these lakes, many of which
are probably much greater than one year (Lin and
Schnoor 1985), the differences in RT between lakes
in Regions 1 and 2 would be greater.
106
Table 5-14. Estimated Hydraulic Residence Time for Drain-
age Lakes and Reservoirs by Subregion (Ex-
cludes Closed and Seepage Lakes), Eastern Lake
Survey-Phase I
Estimated Number
Hydraulic Residence Time (yr)
Subregion
1A
1B
1C
1D
1E
1
2A
2B
2C
20
2
3A
38
and Reservoirs
1091
1421
1276
1071
1429
6288
1105
591
544
2164
4404
250
452
QI
0.06
0.06
0.04
0.03
0.05
0.05
0.15
0.03
0.15
0.14
0.13
0.03
a
Median
0.23
0.25
0.17
0.18
0.23
0.20
0.65.
0.25
0,77
0.49
0.48
0.10
a
o,
0.58
1.03
0.59
0.55
0.75
0.71
2.31
1.47
1.82
1.23
1.63
0.32
a
Estimates not computed
The RT of selected populations of lakes among regions
are compared in Table 5-15. Separate population
estimates of RT were not computed in all regions for
lakes with DOC <2 mg L"1, ANC <0 /yeq L"1, and
reservoirs because of insufficient sample size. Reser-
voirs constitute the primary lake type in Subregion 3A
and most of these were clearwater lakes with DOC
<2 mg L"1. Clearwater lakes in Regions 1 and 2 had RT
Table 6-15. Population Estimates Based on Selected
Characteristics and their Associated Hydraulic
Residence Time, Eastern Lake Survey-Phase I
Hydraulic Residence
Region
1
2
Selected Estimated Numb
Population3 of Lakes
Clearwaterb
Darkwater0
DOCS 2
DOC>6
ANC<0
Reservoirs
Drainage
Clearwater
Darkwater
DOC>6
Drainage
3972
2316
445
1691
263
1216
5072
1601
2803
3065
4222
BP
QI
0.08
0.03
0.09
0.03
0.08
0.04
0.05
0.41
0.07
0.12
0.14
I ill i«7 \yi /
Median
0.31
0.10
0.50
0.09
0.28
0.14
0.23
1.14
0.36
0.39
0.50
0.
1.02
0.33
1.30
0.29
0.72
0.38
0.78
3.54
1.20
1.22
1.63
3A Reservoirs
232
0.03 0.11 0.31
aUnless stated otherwise, populations of drainage lakes and
reservoirs are combined.
bColor <30 PCU
cColor >30 PCU
-------
values approximately three times greater than dark-
water lakes within the same region. The estimates of
RT for low DOC and high DOC lakes parallel the
estimates for the clearwater and darkwater lakes,
respectively. The relationship between DOC and RT is
addressed more fully later in this section. The RT
values for the acidic lakes (ANC<0//eqL~1) in Region
1 were approximately three times greater than the RT
values for populations of darkwater and high DOC
(DOC >6 mg L"1) lakes.
The results presented in Table 5-15 suggest that
there is a strong relationship between DOC (and
related measures such as color) and RT. This rela-
tionship is further described in Figure 5-30, which
shows the Qi and median for classes of RT plotted
against DOC by subregion for the Northeast. The plot
shows an inverse relationship between RT and DOC
for all subregions in Region 1.
The estimated RT values for reservoirs in both Re-
gion 1 and Subregion 3A were very similar and
were exceeded by all selected populations shown
in Table 5-15 with the exception of the darkwater
and high DOC lakes in Region 1. The RT values for
drainage lakes in Region 2 were approximately
2 to 3 times greater than those for drainage lakes in
Region 1. This can be attributed, in part, to the
greater runoff and the higher watershed area:lake
area ratio in Region 1.
In Region 2, lakes with long RT also exhibit the
lowest DOC, but the relationship was less variable
for lakes with RT in the second class (0.5 - 1.0 yr).
The inverse relationship between DOC and RT was
evident in Subregion 3A, but the small concentra-
tions of DOC (generally ^2 mg L~1) in this subre-
gion result in differences <0.5 mg L~1 among the
classes of RT. The lakes with short RT are generally
shallow lakes with large watershed Make area ratios.
This is consistent with Gorham et al. (1985) who
Figure 5-30. Population estimates for median (top bar) and Qi (bottom bar) DOC by three classes of hydraulic residence time
(<0.6,0.6-1.0, and >1 yr, respectively) for lakes in the Northeast (Region 1), Eastern Lake Survey-Phase I.
Hydraulic class
Subregion
107
-------
observed an inverse relationship between color
and maximum lake depth (after excluding lakes
with drainage from peaty areas). They attributed
this relationship to several factors: shallow lakes
are more productive per unit volume than deep
lakes with concomitant greater release of dissolved
organic matter; this greater release of organic mat-
ter to the water column results in interaction with
sediments and overlying water; evaporative concen-
tration per unit volume is greater in shallow than in
deep lakes; and residence times are likejy shorter in
shallow lakes allowing less photooxidation and
microbial decomposition of dissolved organic matter.
The data are not available from this Survey to address
the first three factors, but they do support the
importance of hydraulic residence time in influencing
the DOC concentrations in lakes.
5.4 Characteristics of Acidic lakes
The estimated numbers of acidic lakes (ANC =sO
neq L~1) in most subregions were small with re-
spect to the number required to characterize a pop-
ulation (Section 4.5.1). Acidic lakes were found in
each subregion of the Northeast, although the
number was small in Central New England (35) and
Maine (8). Acidic lakes in the Upper Midwest oc-
curred in the Upper Peninsula of Michigan (2B) and
Northcentral Wisconsin (2C). No acidic lakes were
sampled in the Southern Blue Ridge (3A), but 22
percent of the lakes in Florida (3B) were estimated
to be acidic. The populations of acidic lakes were
too small to characterize adequately for most sub-
regions; therefore, the following discussion of
acidic lakes is based on regional summaries.
The population estimates for selected variables
in acidic lakes are summarized by region/subregion
in Table 5-16. The estimated median sulfate con-
centration (124.7 yueq L"1) in Region 1 is approxi-
mately twice as great as those in Region 2 (61.3 fjeq
L"1) and Florida (3B) (58.8 yueq L"1). The estimated
SO/2 concentrations for the acidic lakes in Region 1
differ little from the estimates for Subregion 1A
(Section 4.8) where most of the acidic lakes in Region
1 are located. The sulfate concentrations for the
acidic lakes in Region 2 differ little from the general
population in the region, but the acidic lakes in Florida
(3B) contain substantially lower concentrations of
SO/"2 than the general population of Florida lakes.
The interquintile differences (Qd) in sulfate concen-
trations for both Region 1 (39.8 jxeq L~1) and Re-
gion 2 (30.8 /ueq L~1) are relatively small, whereas
the interquintile difference is 98.0 ji-eq L~1 for acidic
lakes in Florida. Calcium and extractable Al concen-
trations in Northeastern acidic lakes are also much
greater than those in the Upper Midwest and Florida.
The calcium concentrations in Region 1 acidic lakes
are approximately 18 to 20yueq L"1 greater than those
in Region 2. Over 50 percent of the acidic lakes in
Region 1 have extractable Al concentrations exceed-
ing 100 fjg L~1 whereas less than 3 percent and 12
percent of lakes in Region 2 and Subregion 3B,
respectively, have concentrations exceeding 100 /ug
L~1. The extractable Al concentrations in all three
regions are generally an order of magnitude greater
in acidic lakes compared to concentrations estimated
for the general populations. The estimated DOC
concentrations for the acidic lakes are low (approxi-
mately 50% of the concentrations estimated for the
general populations) and, except for the Qi concen-
tration in Region 2, concentrations are similar among
regions. Low concentrations of DOC in acidic lakes
suggest that most of these lakes are acidic because of
mineral acidity.
The physical characteristics listed in Table 5-16
show that the acidic lakes have a smaller surface
area than the general population of lakes (i.e., all
Table 5-16. Population Estimates of Q, (20th Percentile), Median, and O+ (80th Percentile) of Selected Variables for Acidic
(ANC <0 ixeq L"1) Lakes for Regions 1 and 2, and Subregion 3B, Eastern Lake Survey -Phase I
Variables
Primary
S04-2 (fteq L-1)
Ca+2 (jieq L'1)
Al, ext. (^g L-1)
DOC (mg L'1)
Physical
Lake Area (ha)
Watershed Area (ha)
Site Depth (m)
QI
104.6
33.7
44.8
0.6
7.3
58.0
2.1
1
M
124.7
49.6
105.3
2.1
12.9
146.3
4.3
0,
144.4
63.6
216.2
5.0
34.6
442.2
7.3
QI
48.5
18.0
9.2
2.0
4.7
24.4
1.0
Region
2
M
61.3
31.6
21.8
2.9
6.6
48.1
3.6
04
79.4
43.0
46.7
5.1
12.5
129.4
7.5
QI
29.3
16.5
0
0.6
8.9
46.9
1.7
3B*
M
58.8
30.4
10.7
2.8
14.9
123.6
4.5
04
127.3
45.1
73.9
5.1
35.7
220.3
7.6
*No acidic lakes were sampled in Subregion 3A.
108
-------
acidic and non-acidic lakes) in Region 1 (median
12.9 versus 16.7 ha), and Region 2 (median 6.6 ver-
sus 14.8 ha) and Subregion 3B (14.9 versus 17.3 ha)
(Section 4.4). Estimated values for watershed area
in acidic lakes are much smaller than those for the
general population, which is consistent with an ex-
pected smaller watershed:lake area ratio.
The estimated site depth in acidic lakes for the
Northeast differs little from the estimates for the
general population. The acidic lakes in Region 2 are
considerably shallower than the general population
at all quintiles (median = 3.6 versus 5.6 m, respec-
tively). In contrast, the acidic lakes in Subregion 3B
are deeper than the general population
(median = 4.5 versus 2.7, Q4 = 7.6 versus 5.2 m,
respectively).
109
-------
Section 6
Regional and Subregional Characteristics
6.1 Northeast
Region 1, the Northeast, was estimated to contain
several hundred acidic or low pH lakes (326 with
ANC <0 (xeq L~\ 240 with pH <5.0), more than the
Upper Midwest but not as many as Florida.7 The
Northeast was estimated to have more lakes than
any other region with pH <6.0 (916) or with ANC
<50 (1364) or <200 jjieq l~1 (4258).
Concentrations of sodium and chloride increased
toward the Atlantic Coast; at distances of more
than 50 km from the coast concentrations were low,
with an apparent increase within 20 km. Sodium,
potassium and chloride concentrations in the
Northeast were greater than in the Upper Midwest
but less than in the Southeast. The Northeast had
the highest median sulfate concentration (115.4
(jieq L~1) of all regions. The median calcium concen-
tration (177.4 jjieq L~1) was lower than that in the
Upper Midwest or Florida, and concentrations of
magnesium were the lowest among the regions.
However, only 5 percent of lakes in the Northeast
had calcium <50 jjieq L"1, which was the lowest
percentage of lakes below this reference value for
all regions. Unlike other regions, acidic lakes in the
Northeast were characterized by high levels of ex-
tractable Al (for lakes with ANC <0 yueq L~1, the
estimated median extractable Al was 1 05.3 /ug L~1 for
the Northeast, 21 .8 /ug L~1 in the Upper Midwest, and
10.7 fjg L"1 in Florida). Acidic lakes in the Northeast
also had higher concentrations of calcium, mag-
nesium and sulfate than acidic lakes in other regions.
The rank order of concentrations of major cations
and anions in subregions of the Northeast at both
the Q1 and median population values were dis-
cussed in Section 5.2.5.2 (Table 5-5). In Subregion
1D, Na+ was the dominant cation. In Subregions
1 A, 1 B, and 1 C, SO/2 was the dominant anion at Qi;
at the median value, however, S04~2 was dominant
only in Subregion 1 A, with HCOa" dominant in Sub-
regions 1B, 1C and 1E. Chloride dominated at both
7AII references to medians, quintiles, percentages, numbers, or areas are
population estimates but are not always noted as estimates in the text.
Upper confidence limit information is provided in the tables but is not
discussed in the text.
estimated values in Subregion 1D. Using the mean
from 79 lakes in New England, Brooks and Deevey
(1963) found a rank order of HC03" > S04"2 > CI"
and observed that most of the lakes were bicarbon-
ate systems.
When unmeasured anion concentrations (A";
Table 5-6) were included as a measure of organic
anion, the rank order of anions changed little (Sec-
tion 5.2.5.2, Table 5-6) in Region 1. At Q, levels, the
anion deficit was zero for Subregions 1A, 1B and
1D. Anion deficit was only fourth in importance in
Subregion 1C and third in Subregion 1E after bicar-
bonate and sulfate. At median values, the anion
deficit was also zero in Subregion 1D. It was fourth
in Subregions 1B and 1C and third in Subregions 1A
(S04~2 > HCO3~ > A') and 1E (HC03~ > S04"2 > A").
Medians and interquintile differences (Q4 - Q,
= Qd) are given for several variables in Table 6-1.
The interquintile difference (Qd) is a measure of the
variability of chemistry among lakes within subre-
gions. For ANC, calcium, and sulfate, Maine (1E)
showed less variability than the other subregions in
the Northeast, and the Poconos/Catski!ls(1 B) showed
the greatest variability.
The Northeast was the only region where most
lakes (65%) were clearwater (<30 PCU). Dissolved
organic carbon concentrations in darkwater (>30
PCU) northeastern lakes were much lower than
those in darkwater lakes in the Upper Midwest or
Florida. Total phosphorus and silica concentrations
were higher in darkwater lakes than in clearwater
lakes. Darkwater lakes were also shallower and
smaller, but their watershed areas were often larger
than those of clearwater lakes.
Seepage lakes were not common in the Northeast,
representing only 7 percent of the lakes classified
by hydrologic type. Drainage lakes, 71 percent of
the total, were the most common hydrologic type,
followed by reservoirs (17%). Seepage lakes gener-
ally had the lowest values of ANC and pH of any
lake type. The median ANC for seepage lakes was
84 |xeq L"1 versus 145 (xeq L"1 for drainage lakes. In
seepage lakes the median pH was 6.5 and in
drainage lakes, 6.8. These observations of lower
110
-------
Table 6-1.
Median* and Interquintile Difference* (Q4 - Qi = Qd) for pH, ANC, Calcium, Sulfate, and DOC, Eactern Lake
Survey-Phase I
pH
ANC
Ca
+2
DOC
Region
and
Subregion
1A
1B
1C
1D
1E
1
2A
2B
2C
2D
2
3A
3B
Median
6.71
7.02
6.77
6.81
6.91
6.87
6.94
7.10
6.68
7.39
7.09
6.98
6.56
Qd
1.66
0.72
1.00
0.93
0.74
0.93
0.77
1.75
1.72
1.44
1.46
0.56
2.03
Median
111.8
297.4
119.9
161.8
148.4
158.1
184.9
283.6
93.9
801.7
359.5
250.2
83.5
Qd
236.8
428.3
335.7
363.2
226.3
348.3
305.2
1228.9
584.9
1917.9
1336.7
272.2
720.5
Median
144
291
137
187
147
177
143
246
101
522
238
105
238
Qd
201
344
314
296
191
293
232
883
346
1102
856
184
799
Median
119
159
101
141
75
115
62
78
57
50
57
32
94
Qd
38
103
58
82
31
86
37
54
34
56
53
40
244
Median
4.1
3.8
4.4
4.0
5.2
4.3
9.2
6.8
4.6
8.8
7.5
1.9
8.6
Qd
3.0
3.1
3.8
4.5
5.5
3.8
7.2
7.8
4.9
7.8
7.7
2.8
9.6
ANC and pH in seepage lakes in Region 1 are con-
sistent with the pattern found in the Upper Mid-
west.
Nutrient concentrations were low for most of the
lakes in the Northeast. Median concentrations were
estimated to be 0.4 jjt-eq L~1 nitrate, 1.4 |xeq L~1
ammonium and 9.0 jig L~1 total phosphorus. True
color values were similar among subregions, and
turbidity was generally low, resulting in a median
Secchi disk transparency of 2.3 m. DIG values were
low (2.1 mg L~1 estimated median).
6.1.1 Adirondacks (1A)
The Adirondacks contained the highest percentage
and number of acidic lakes (ANC sfl jjieq L~1) and
lakes with pH <5.0, of any subregion except Florida
(3B). A total of 138 lakes, representing 11 percent of
the target population, was estimated to have ANC
<0 n-eq L~1. A similar number of lakes (128) was
estimated to have pH values <5.0. Compared with
other areas in the Northeast, Subregion 1A had
higher percentages and numbers of lakes with ANC
values <50 jxeq L~1 (36% or 459 lakes) and pH <6.0
(27% or 343 lakes).
Far more lakes (82) in the Adirondacks, classified as
clearwater (<30 PCD), had extractable aluminum
concentrations >150 (ieq L~1 than any other subre-
gion sampled as part of the ELS-I. Only 26 such
lakes occurred in all other subregions combined.
Other studies have also found elevated aluminum
associated with low pH in Adirondack lakes
(Driscoll et al. 1980, Schofield 1976, 1982).
The median sulfate concentration in Subregion 1A
was 118.7 jxeq I"1. This value was lower than those
observed in Southern New England (10) and the
Poconos/Catskills (1B), but was higher than in Cen-
tral New England (1C) and Maine (1E). Thirteen per-
cent of the lakes in the Adirondacks had sulfate
concentrations a150 n-eq L~1 compared to 56 per-
cent in 1B and 46 percent in 1D. The population
estimates for the concentrations of sulfate in Subre-
gion 1A are similar to results of a survey of 214 high
elevation lakes in the Adirondacks done by
Schofield (1976) in 1975. The mean sulfate concen-
tration for that survey was 133 jjteq L~1. The Adiron-
dacks contained the largest surface area of lakes
with high sulfate in the Northeast. This is because
several large lakes in the western portion of this
subregion had high sulfate concentrations.
Sodium and chloride concentrations in the Adiron-
dacks were the lowest in the Northeast. This subre-
gion also had the lowest first quintile magnesium
concentration in the Northeast. In the Adirondacks,
calcium was the dominant cation and sulfate was
the dominant anion. The rank order of ions ob-
served in ELS-I was consistent with other observa-
tions for the area (Peters et al. 1981, Schofield
1982).
Specific subpopulations were analyzed for the east-
ern and western areas of the Adirondacks, using
lakes located on either side of a meridian of 74° 30'
00". Whereas sulfate concentrations were the same
on the western (windward) and eastern (leeward)
areas of the Adirondack drainage divide, lakewater
pH, ANC and calcium were estimated to be lower in
the west. Median ANC was 39.3 ixeq L~1 west of the
divide and 166.7 ^eq L~1 east of the divide. Median
calcium was 101.9 (xeq L~1 and pH was 6.15 on the
west side as compared to 202.5 (xeq L"1 and 6.88,
respectively, on the east side. The percentage of
111
-------
lakes with calcium <50 fieq Lr1 was 12.5 percent
versus 4.0 percent on the west and east sides, re-
spectively. Correspondingly, median extractable
aluminum was also higher on the windward side
(12 versus 5 fig L~1). Dissolved organic carbon con-
centrations were similar in both areas (d, = 2.4 mg
L~1 in the west and 2.9 mg L~1 in the east, median
= 4.3 mg L~1 in the west and 3.9 mg L~1 in the east).
Consequently, even though sulfate concentrations
were the same on both sides, the lower values of
calcium at all quintiles on the west side resulted in
more acidic lakes.
Population estimates for lakes having ANC =£0 fieq
L~1 were significantly higher on the west side (20%,
130) than on the east side (1.3%, 8). Similar esti-
mates were obtained using pH =£5.0 as a reference
value (120 in the west and 8 in the east). The pro-
portion of lakes with calcium <50 fieq Lr1 also
varied (12.5% in the west and 4.1% in the east). Only
1.3 percent of the lakes on the east side were esti-
mated to have extractable aluminum £50 fig L~1,
whereas 29 percent of the lakes on the west side
had values >50 fig L~1 and 17 percent had values
^150 jig L~1. Values of total aluminum were also
high in the Adirondacks. In addition, both pH and
ANC were negatively correlated with elevation in
the Adirondacks. This relationship was not ob-
served in any other area in the Northeast. Likewise,
Haines and Akielaszek (1983) found only a weak
relationship between elevation and ANC in clear-
water, headwater lakes in New England.
Most of the lakes in the Adirondacks were clear-
water. Lakes that had pH values £5.0 were equally
likely to be darkwater (34%) as were those with
higher pH. Dissolved organic carbon concentra-
tions in darkwater lakes were relatively low com-
pared to other areas; only 39 percent of the dark-
water systems had DOC 2=6 mg L~1. Darkwater
lakes in the Adirondacks tended to be smaller, but
had larger relative watershed areas than the clear-
water lakes. The watershed area:lake area ratios
were about 20 for darkwater, and 10 for clear-water
systems.
Total phosphorus in the Adirondacks was the low-
est in the Northeast (Q, = 1.6 fig L~1, M = 4.7 fig
L~1). Together with low true color and DOC, this
resulted in the highest Secchi disk transparencies in
the Northeast (M = 2.8 m, Q4 = 4.9 m).
Drainage lakes were the most common lake type.
This hydrologic type was estimated to represent 77
percent of the population. Drainage lakes were
characterized by low ANC and pH values as well as
the highest extractable aluminum concentrations.
It is important to point out that the Adirondacks
contain many lakes <4 ha in size, as do some other
areas. Lakes <4 ha were not included on all small-
scale maps used for ELS-I lake selection; therefore,
in developing the frame population, it was neces-
sary to exclude lakes <4 ha from sampling to estab-
lish consistency and reliability in estimating the
target population. The chemical status of lakes
<4 ha is unknown from the ELS-I data, but lakes
<4 ha are being sampled by the Adirondack Lake
Survey Corporation (ALSC; Colquhoun et al. 1984).
Results from ELS-I cannot be extrapolated to lakes
smaller than 4 ha and their chemistry cannot be
inferred from ELS-I measurements on larger lakes.
Comparison of the chemistry of ALSC lakes <4 ha
to ELS-I lakes in the Adirondacks would provide
insight into the importance of small lakes in esti-
mating the proportion of acidic and low pH lakes in
the total population of lakes in the Adirondacks.
6.1.2 Poconos and Catskills (1B)
The Poconos/Catskills contained few low pH lakes
or acidic lakes. Twelve lakes (1%) were estimated to
have pH values <5.0 and 78 lakes (5%) had ANC sO
(ieq L~1. In contrast to other areas surveyed, acidic
lakes were larger than non-acidic lakes. As a result,
when the area of acidic lakes is compared to other
subregions, the Poconos/Catskills contain the
largest surface area of acidic lakes in the Northeast
(2937 ha). This subregion had 116 lakes with pH
<6.0 and 194 lakes with ANC <50 (ieq L~1. It was,
however, the only part of the Northeast where less
than half of the lakes had ANC values <200 fieq L~1.
Median calcium concentrations in Subregion 1B
were the highest in the Northeast. Magnesium con-
centrations were also high. Sulfate concentrations
in the Poconos/Catskills were very high. Median
and QT sulfate concentrations were the highest in
the region (Q-, = 118.1 fieq L~1, M = 159.3 fieq L~1).
Few of the lakes had sulfate concentrations <50
(ieq L~1. Similarly high concentrations of sulfate
were found by Bradt and Berg (1983) and Bradt et
al. (1984). The fourth quintile value for extractable
aluminum in Subregion 1B was the lowest of ail
subregions in the Northeast, as was total alu-
minum. Silica concentrations were low, but nutri-
ents were the highest in the Northeast. Lakes in the
Poconos/Catskills were clearly different in nearly all
aspects from other areas of the Northeast. Reser-
voirs were common (42%) and the median water-
shed area was 169 ha.
6.1.3 Central New England (1C)
Few acidic or low pH lakes were found in Central
New England (an estimated 2 percent or 25 lakes
with pH <5.0, 35 with ANC <0 fieq L~1). Smaller
lakes were generally more acidic. No clearwater
lakes had extractable aluminum concentrations
>100 fieq L"1. Shallow lakes in Central New Eng-
land generally had higher ANC than deep lakes.
Central New England had 191 lakes with pH <6.0
112
-------
and 262 lakes with ANC <50 |ieq L"1. This subre-
gion also had the second highest percentage of
number (68%) and the highest percentage of area
(74%) of lakes with ANC <200 ji,eq L~1 in the North-
east. Subregion 1C had few (63) lakes with calcium
concentrations ^50 jxeq L"1, but the lowest median
calcium concentration (136.8 n-eq L~1) in the North-
east.
Most of the lakes (68%) in Central New England
were clearwater, with high Secchi disk transparen-
cies. Unlike other northeastern subregions, dark-
water lakes in Central New England had median
and quintile pH and ANC values that were lower
than those for clearwater lakes. Central New Eng-
land lakes were predominantly drainage lakes
(81%). The seepage lakes, representing 7 percent of
the lakes in this subregion, generally had lower
ANC and pH and higher extractable aluminum than
drainage lakes. This was also similar to observa-
tions in Subregion 2C.
Within the subregion, marked differences were
noted between lakes in Vermont and New Hamp-
shire. None of the lakes in Vermont was estimated
to have ANC <0 |xeq Lr1, but 17 lakes in New Hamp-
shire were estimated to have ANC ^0 jteq L~1. Also,
many more lakes in New Hampshire (537) were es-
timated to have ANC <200 jieq L"1 than in Vermont
(90). This is consistent with maps of bedrock geol-
ogy (Norton 1982, Hendrey et al. 1980b), descrip-
tions of geologic substrates by Brooks and Deevey
(1963) and mapping of lake alkalinities by Clarkson
(1982). Lakes with lower ANC values were clustered
in northern and southern Vermont, with higher
ANC lakes in the remainder of the state. Lakes with
higher values of ANC were found particularly in the
Champlain Lowland. A similar pattern was noted by
Haines and Akielaszek (1983). Norton et al. (1981)
also observed clearwater, low ANC lakes in New
Hampshire. In contrast to Vermont, low ANC (<50
ixeq L~1) lakes were distributed throughout New
Hampshire, and as noted, New Hampshire was esti-
mated to have many more lakes with ANC s=200
(xeq L~1 than Vermont. Lakes with ANC <200 jxeq
L~1 were clustered in southeastern New Hampshire
and along the border with Vermont (Section 4.5.1.2,
Figure 4-11).
6.1.4 Southern New England (ID)
Five percent of the lakes in Southern New England
were acidic and 5 percent had low pH (66 lakes with
ANC <0 jjueq L~1, 66 lakes with pH <5.0). Although
both pH and ANC showed positive relationships
with lake size, this subregion contained the largest
lake area with low pH in the Northeast (6% or 2295
ha with pH <5.0). Southern New England had 192
lakes with pH <6.0, 284 with ANC <50 jieq L~1, and
755 with ANC <200 fieq L'1. Lakes with low ANC
were not distributed uniformly, with higher values
found in southern Connecticut and parts of Massa-
chusetts (Section 4.5.1.2, Figure 4-11).
Concentrations of sulfate were the second highest
in the Northeast, with a median value of 141.1 fj,eq
L~1. Subregion 1D contained more lakes (133) with
calcium <50 jxeq L~1 than any other northeastern
subregion. Median sodium and chloride concentra-
tions in Southern New England were the highest of
any subregion, and median potassium and mag-
nesium concentrations were the highest in the
Northeast. High sea-salt contributions were also
observed by Haines and Akielaszek (1983), who
observed a similar decline in chloride concentrations
with distance from the coast.
The majority (62%) of Southern New England lakes
were clearwater. Southern New England had a
much higher percentage (18%) of lakes with low
DOC concentrations (<2 mg L~1) than any other
northeastern subregion. Unlike other subregions of
the Northeast, darkwater lakes had much higher
calcium concentrations than clearwater lakes.
Drainage lakes were predominant (52%) in South-
ern New England, although reservoirs were also
important (29%). Seepage lakes generally had the
lowest pH and ANC of any lake type although they
comprised only 13 percent of the lake population.
Closed lakes had the highest median extractable
alumirium concentration. Lakes in Southern New
England showed the greatest variation in elevation
of any subregion in the Northeast (Q4 - 0.,
= 567 m) and were also the shallowest (Q4= 5.9 m)
of any subregion except Florida. Subregion 1D con-
tained the most small watersheds in the Northeast
(QT = 65 ha). Silica concentrations in this subregion
were relatively high for the Northeast (M = 2.1 mgL~1),
but variable (Qd = 4.3). Phosphorus concentrations
were also relatively high compared to other sub-
regions (M = 14.0 A»gL~1), similar to values reported by
Deevey (1940) for Connecticut (M = 14.9 //g L"1).
6.7.5 Maine (IE)
Maine had the fewest acidic or low pH lakes in the
Northeast. Only one half of one percent of the lakes
had pH <5.0 or ANC <0 jieq L~1. Maine also had the
fewest lakes with pH <6.0 (5%) or ANC <50 jieq L~1
(11%). However, Maine had the most lakes with
ANC <200 jxeq L~1 (1020 or 67%) in the Northeast,
comprising 109,546 ha. Shallow lakes generally had
higher ANC than did deeper lakes. Lakes with ANC
<200 |j,eq L~1 were found throughout Maine (Sec-
tion 4.5.1.2, Figure 4-11). ANC values >200 jieq L~1
were found in several areas, but principally in the
northern third of the state. The spatial distribution
of lakes with high ANC was similar to that found by
Mairs (1966) and Haines and Akielaszek (1983).
113
-------
Maine lakes had calcium concentrations that were
intermediate in the Northeast. None of the clear-
water lakes sampled in Maine had extractable alu-
minum >50 |xg L~1, in sharp contrast to the Adiron-
dacks, where 14 percent had extractable aluminum
==50 jxg L~1. Maine had the lowest median sulfate in
the Northeast (74.6 jxeq L~1). Only one percent of
the lakes in Maine had sulfate concentrations >150
(jieq L~1. Other surveys have also found lower sul-
fate values in Maine than in other areas of the
Northeast (Norton et al. 1981, Haines and
Akielaszek 1983). Maine was the only area in the
Northeast where HC03" > SO^2 at Qi levels. Chlo-
ride concentrations declined sharply with distance
from the coast, as other studies (Mairs 1967, Haines
and Akielaszek 1983, Norton et al. 1981} have
shown. Concentrations of calcium were higher than
those of other cations.
Maine contained the lowest percentage of clear-
water lakes (55%) and had a much higher percent-
age (42%) of lakes with DOC >6 mg L~1 than any
other northeastern subregion. Others have noted
the significant number of lakes with perceptible
color (Davis et al. 1978b). Maine was unique in
the Northeast in having darkwater lakes with higher
ANC concentrations than clearwater lakes for the
median and quintile values. As in the Adirondacks,
darkwater lakes in Maine generally were smaller
but had larger watersheds than clearwater lakes
(median WA:LA was 23 for darkwater lakes, 9 for
clearwater lakes). Darkwater lakes in Maine also
were shallower than clearwater lakes. Darkwater
lakes had high DOC concentrations, similar to ob-
servations in the Upper Midwest and Florida.
Maine had the highest percentage of drainage lakes
in the Northeast (91%). The largest median lake
area (29.1 ha), and watershed area (472 ha), and
some of the deepest lakes (Q4 = 10.6 m) also were
found in Maine. Nutrient concentrations were low
(estimated median total phosphorous = 5.7 ng L~1)
and Secchi disk transparency values high (M
= 2.6 m, Q4 = 4.7 rn), consistent with other investi-
gations (Cooper 1942, Bailey 1975, Davis et al.
1978b).
6.2 Upper Midwest
The Upper Midwest shows considerable hetero-
geneity in lake chemistry within subregions (Table
6-1) and major differences in chemical and physical
characteristics of lakes among subregions. As an
example, lakes in Subregions 2B and 2D exhibited
the largest interquintile differences (Qd) for ANC
and calcium. Interquintile differences for the other
variables were exceeded only by those in Subre-
gion 3B. The contrasts within Region 2 are evident
by comparing the low ANC lakes of Northcentral
Wisconsin (M = 93.9 jxeq L~1) with those of the Up-
114
per Great Lakes Area (M = 801.7 n,eq L 1).
The Upper Midwest was estimated to have only 148
lakes with ANC =£0 jjieq L~1, far fewer than the esti-
mate for the Northeast and Subregion 3B. How-
ever, the estimated number of lakes with ANC =£50
(jieq L~1 in Region 2 (1312) was comparable to that
in Region 1 (1364) and nearly twice as great as in
Subregion 3B (742).
Lakes in Region 2 were characterized by low con-
centrations of many constituents including ex-
tractable aluminum, sodium, sulfate and chloride.
Dissolved organic carbon concentrations in Re-
gion 2 (M = 7.5 mg L~1) were almost twice as high
as those in Region 1 (M = 4.3 mg L~1); an estimated
63 percent of the lakes had DOC concentrations s=6
mg L"1.
The order of major cations in the Upper Midwest
was typical of ion orders observed in other calcium
bicarbonate lakes. The order of major anions at the
median was generally typical of bicarbonate lakes
(HC03~ > S04"2 > Cl~) although lakes in Subregions
2B and 2C had relatively high concentrations of sul-
fate at Qv Including organic anions (calculated
from the anion deficit) in the order of anions illus-
trated the potentially important influence of organic
anions on lake chemistry in the Upper Midwest
(Section 5.2.5.2, Table 5-6).
6.2.1 Northeastern Minnesota (2A)
Most of the lakes in Northeastern Minnesota had
relatively high ANC and pH. No acidic (ANC <0 /ueq
L"1) lakes were sampled in Subregion 2A, a finding
consistent with the results of a sampling of 290
Northeastern Minnesota lakes reported by Schnooret
al. (1986). Only 1 percent of the lakes in the subregion
were estimated to have a pH <6.0. Calcium concen-
trations were also high; the Q; value for this variable
was the highest in the Upper Midwest. Sulfate
concentrations were low in Northeastern Minnesota
lakes (0.4 = 83.9 fjeq L~1), but were intermediate
compared to other subregions in the Upper Midwest.
The most striking feature of the lakes in Northeast-
ern Minnesota was their high color (70% are con-
sidered darkwater) and DOC (76% of the lakes have
DOC 2:6 mg L~1) values. This apparently results
from the extensive peatland and the connected
drainage lake networks (74% of the lakes are
drainage) (Rapp et al. 1985; Omernik and Griffith
1985). The WA:LA ratio for darkwater lakes in Sub-
region 2A was 13.3, but this variable alone is insuf-
ficient to explain the high color and high DOC in
these lakes.
The lake chemistry in Northeastern Minnesota
showed relatively little heterogeneity. For example,
the Qd for pH in Subregion 2A was 0.77 pH units
-------
compared to 1.75, 1.72 and 1.44 for Subregions 2B,
2C and 2D, respectively (Table 6-1). A similar pat-
tern was observed for ANC where Qd was 305.2 for
Subregion2Aand 1228.9,584.9 and 1917.9 for Sub-
regions 2B, 2C and 2D, respectively.
The lakes in Northeastern Minnesota had slightly
larger lake surface and watershed areas than other
lakes in the Upper Midwest, but the WA:LA ratios
were not substantially greater than those of other
subregions. Consequently, the higher ANC concen-
trations in the Northeastern Minnesota lakes may
be attributed, in part, to the composition of the sur-
ficial and bedrock geology. Other studies of lake
chemistry and watershed factors in Minnesota
(most of these studies extending beyond the
boundaries of Subregion 2A) show an association
between soils (Gorham et al. 1983), glacial till
(Moyle 1956) and bedrock geology (Rapp et al.
1985). However, Winter (1977) considered ground-
water chemistry to be a better index of lake chem-
istry because it incorporates hydrology with the ge-
ology. The moderately high sulfate concentrations
in Northeastern Minnesota lakes are also likely to
result from the presence of sulfide ores in the
watershed (Rapp et al. 1985). The extractable alu-
minum concentrations in clearwater lakes for Sub-
region 2A were low, and none of the lakes sampled
had values >100 |xg L~1. Most of the lakes with high
aluminum values in Subregion 2A also had high
DOC. It is assumed that most of the extractable
aluminum in this case is complexed with organic
ligands (Driscoli et al. 1980).
6.2.2 Upper Peninsula of Michigan (28}
The Upper Peninsula of Michigan was estimated to
have the highest percentage of acidic lakes (10%) in
the Upper Midwest. An estimated nine percent of
the lakes in this subregion had pH <5, also the
highest percentage in the Upper Midwest. How-
ever, because there is a relatively large population
of high ANC lakes in Subregion 2B, median values
for pH (7.10) and ANC (283.6 (jueq L~1) are high rela-
tive to other subregions. The acidic lakes (ANC <0
pieq L~1) in Subregion 2B were generally clearwater
(median color = 22 PCU). Thirty-eight percent of
the lakes in Subregion 2B were seepage lakes with
very low concentrations of most ions (median Ca+2
= 111 fjieq L"1; SO4"2 = 67 /ueq L"1). Drainage lakes
comprised 51 percent of the lake type. Sulfate con-
centrations were low (Q4 = 103.5 jxeq L~1) in Subre-
gion 2B compared to lakes in Region 1, but they
were approximately 20 percent greater than in
other subregions in Region 2. Silica concentrations
were extremely low (Q, for silica in Subregion 2B
lakes was 0.3 mg L"1). The extractable aluminum
values were generally low (only 0.4% of the clear-
water lakes had values >100 fj,g L~1), although the
Q4 value of 11.9 fig L~1 was relatively high.
Large interquintile differences for all primary vari-
ables except sulfate were observed. This hetero-
geneity in Subregion 2B can be explained by the
contrasting chemical composition of the bedrock
types and the presence of high percentages of both
seepage (38%) and drainage (51%) lakes (Schnoor
et al. 1986). Subregion 2B also represents a large
geographic area, extending approximately 250 km
from east to west, and exhibits a considerable gra-
dient in deposition chemistry (Glass and Loucks
1986). The greater numbers of Jow ANC lakes ob-
served in the eastern portion of the peninsula (Sec-
tion 4.5.1.3, Figure 4-12) are consistent with the pat-
tern observed by Schneider (1975). The number of
acidic lakes in the eastern portion (east of longitude
87°) of the Upper Peninsula of Michigan was esti-
mated to be 18.2 percent, which also agrees with
the estimate of 19 percent by Schnoor et al. (1986).
6.2.3 Northcentral Wisconsin (2C)
Northcentral Wisconsin, although comprising the
smallest land surface area of all subregions, had an
estimated target population of 1480 lakes, exclud-
ing those >2000 ha. Whereas only a small percent-
age (3%) were estimated to be acidic (ANC <0 |xeq
L"1), Subregion 2C had the highest percentage
(41%) of lakes among all subregions with ANC <50
jxeq L"1. These results agree closely with the find-
ings of Eilers et al. (1983) and Schnoor et al. (1985)
who reported 3 percent and 4 percent, respectively,
of lakes in this area with ANC <0 (j,eq L~1. The
median ANC for lakes in Northcentral Wisconsin
(93.9 (xeq L~1) was the lowest among all subregions
except Florida (83.5 n,eq L~1).
The factors believed to contribute most to the low
ionic concentrations of lakes in Northcentral Wis-
consin are hydrology (Juday et al. 1935, Broughton
1941, Eilers et al. 1983, Lin and Schnoor 1986), soils
(Gorham et al. 1983), and composition of the glacial
till (Simpkins et al. 1978). Seepage and drainage
lakes exhibited markedly different chemistry. The
median WA:LA ratio for lakes in Subregion 2C (6.1)
was lower than for other lakes in the Upper Mid-
west and most other lakes in the Survey (with the
exception of Subregion 3B, 5.5). The modest influ-
ence of watershed factors results in low weathering
rates reflected in low concentrations of silica and
calcium (Schnoor et al. 1986). The acidic lakes in
Northcentral Wisconsin are similar in many chemi-
cal aspects and physical features to those in the
Upper Peninsula of Michigan and Florida (i.e., clear-
water, low ionic concentration, seepage lakes).
6.2.4 Upper Great Lakes Area (2D)
No acidic (ANC <0//eq L"1) lakes were sampled in the
Upper Great Lakes Area. The subregion had a median
ANC of 801.7 /jeq L"1, by far the highest among all
subregions. It also contained the largest number of
115
-------
lakes (1411) with ANC<200/jeqL 1ofanysubregion.
An estimated 20 percent of the lakes had pH values
>8.07; 4 percent of the lakes had pH values <6.0. The
proportion of clearwater and darkwater lakes in
Subregion 20 was approximately equal. Lakes in
Subregion 2D contained high concentrations of many
major ions (Sections 4.7 a nd 4.8, Tables 4-25 through
4-30), but the concentration of extractable aluminum
was relatively low in clearwater lakes (Q« = 8.2
Most of the lakes in Subregion 2D with ANC =£50
jjieq L~1 are located in Northwestern Wisconisn
(Section 4.5.1.3, Figure 4-12). This agrees with the
expected pattern based on Omernik and Griffith
(1985) and a survey by Lillie and Mason (1983).
Subregion 2D contains at least three distinct geo-
logical provinces that result in strongly contrasting
chemistry among lakes in Minnesota, Wisconsin
and Michigan. The Minnesota portion of Subregion
2D was described by Moyle (1956) as a "hard-water
flora" area, which was attributed, in part, to the
presence of Paleozoic and Cretaceous sedimentary
rocks. Carbonate in glacial till (Brugam 1981) and
the soils (Gorham et al. 1983) in this part of Minne-
sota may also contribute to high concentrations of
calcium carbonate in these lakes. The lower penin-
sula of Michigan, also contained within Subregion
2D, is underlain by Cretaceous sedimentary carbon-
ate rocks (Barr 1978); consequently, lakes in this
area also had high values of ANC. High ANC values
in lower Michigan are in agreement with the study
by Schneider (1975). In contrast, the Wisconsin por-
tion of 2D is underlain by sedimentary sandstones
and igneous bedrock (Mudrey et al. 1982) and cov-
ered by 20 to 200 m of glacial till containing a small
percentage of carbonate till (Hadley and Pelham
1976).
6.3 Southeast
Region 3, the Southeast, contains only two subre-
gions. As mentioned previously, these two disjunct
subregions contain lakes that are extremely differ-
ent in physical and chemical characteristics and will
be discussed separately.
6.3.1 Southern Blue Ridge (3A)
No acidic or low pH lakes were sampled in the
Southern Blue Ridge. This subregion had the low-
est percentage of lakes with ANC <50 p,eq L~1 (1%)
or pH <6.0 (<1%) of any subregion. It also con-
tained the second lowest percentage of lakes for
any subregion with ANC <200 n-eq |_~1 (34%). ANC
and pH values were generally lowest in the large,
deep lakes.
The Southern Blue Ridge had the lowest median
sulfate concentration (31.8|j,eq L~1) and the second
lowest median calcium concentration (104.7
L~1) of any subregion. No clearwater lakes with ex-
tractable aluminum concentrations s50 jxg L~1
were sampled.
The majority of the lakes in the Southern Blue Ridge
were classified as darkwater (55%), but an evalua-
tion of DOC and anion deficit indicates that the
color is not related to organic matter. Apparently
inorganic contributions to color were important in
these lakes; only 11 percent of the darkwater lakes
had DOC a6 mg L~1, which was far less than in any
other subregion. Southern Blue Ridge lakes had the
lowest DOC concentrations of any subregion (54%
had DOC <2 mg L~1 and only 6% had DOC >6 mg
LT1). Darkwater lakes in the Southern Blue Ridge
had smaller median lake and watershed areas than
clearwater lakes but were similar in other character-
istics.
An estimated 90 percent of the lakes were reser-
voirs. As is typical for reservoirs, lakes in this subre-
gion had the largest median watershed area
(682 ha), smallest median lake area (10.8 ha) and
the highest median turbidity (3.9 NTU) of any sub-
region (Thorton et al. 1980). The Southern Blue
Ridge had the highest median silica (9.0 mg L~1) and
total aluminum (81.7 jug L"1) concentrations of any
subregion, even though calcium and sulfate were
low. Median nitrate (3.1 ^eq L"1) and potassium (39.4
/aeq L~1) were higher than in all other subregions.
Such high constituent concentrations are typical of
Southern Blue Ridge reservoirs (Placke 1983).
6.3.2 Florida (3B)
Florida contained the largest number of acidic or low
pH lakes of any subregion (22% or 463 lakes with ANC
<0//eq L"1; 12% or 259 lakes with pH <5.0). Previous
limnologies! studies have indicated that many lakes
in Florida could generally be classified as acidic,
softwater systems (Shannon and Brezonik 1972;
Baker et al. 1981; Canfield 1981). Acidic lakes had
somewhat higher extractable aluminum concentra-
tions than the subregion as a whole, but very few
Florida lakes had high concentrations of extractable
aluminum. This is consistent with the findings of
Hendry and Brezonik (1984), who noted acidic lakes
in Florida do not currently have biologically mean-
ingful aluminum concentrations. An estimated four
percent of the clearwater lakes had extractable
aluminum >100 //g L"1. This may be because
aluminum concentrations are low in the sandy soils
of Florida where many of the acidic lakes are located
(Brezonik et al. 1983). However, aluminum in the
sediments of Florida lakes may become available for
release to overlying waters as pH of the water
declines (Baker, L 1984).
116
-------
Seepage lakes were the most common type (66%)
in Florida, followed by drainage lakes (22%) and
closed lakes (13%). Florida lakes were generally
medium-sized and shallow, and with small water-
sheds. Seepage lakes in Florida had the smallest
lake and watershed areas but were deep. They were
generally lower in ANC, pH, color, sulfate and cal-
cium than drainage lakes. Sodium and chloride
concentrations were high in Florida lakes; silica
concentrations were exceptionally low (M = 0.3 mg
L"1). The order of cations and anions (Section 5.2.5.2,
Table 5-5) was similar to that in Subregion 1D. At the
Qi value, Na+ and Cl~ were dominant.
The highest variability observed for pH, sulfate and
DOC occurred within Subregion 3B (Table 6-1).
Based on the sample data, there appear to be at
least five areas within the Subregion containing
lakes with distinct water chemistry characteristics.
These are Southeast Georgia or the Okefenokee
area, the Highland area of the Florida Panhandle,
the Northcentral Peninsula of Florida, the Gulf
Coast of Florida, and the Southcentra! Peninsula of
Florida. Selected subpopulation estimates were
computed for these areas and will be discussed in
this section. The boundaries of these areas are
shown in Figure 6-1.
In Florida, 40 percent of the lakes had sulfate con-
centrations >150 jjieq L"1. In the Panhandle, 75 per-
cent of the lakes (195) were estimated to have ANC
==0 (ieq L~1 while in the Northcentral Peninsula only
21.7 percent of the lakes (209) had ANC sO fieq L"1.
In the Panhandle, only 15.2 percent of the lake pop-
ulation had SO«~2 >50 /ueq L"1 while in the North-
central Peninsula 91.9 percent had S04"2 >50/L»eq L~1
and 60.7 percent had SO*'2 >150 //eq L"1.
The majority of lakes in Subregion 3B were dark-
water (56%). Unlike the reservoirs of the Southern
Blue Ridge, color and DOC levels were related (for
example, all darkwater lakes had DOC >6 mg L"1).
DOC was generally high in Florida lakes (69% had
DOC >6 mg L'1). Clearwater lakes in Florida were
deeper, and had lower ANC and pH values, and
calcium concentrations than did darkwater lakes.
Canfield et al. (1984) found that color increased
from north to south and inland from the coast. The
subpopulation of lakes in the Okefenokee area (Fig-
ure 6-1) of 3B were highly colored and contained no
lakes with DOC <6 mg L"1. In the Panhandle, 65
percent of the lakes had DOC >6 mg L"1 and the
Northcentral Peninsula had 31 percent.
The Okefenokee and the Florida Panhandle subpopu-
lations both had lakes with small surface areas (M =
6.1 and 10.2 ha, respectively), low pH (M = 4.1 and
4.9, respectively), low calcium (M = 17.9 and 18.4 /ueq
L"1, respectively) low sulfate (M = 4.7 and 31.9 ,ueq
L'1, respectively) and mostly negative ANC values (M
= -127.2 and -23.7 /aeq L"1, respectively). The Florida
Panhandle lakes had low color (M = 14.7 PCU) and low
DOC (M = 3.9 mg L"1) and silica concentrations (M =
0.07 //eq L"1), whereas the Okefenokee lakes were
very shallow (M = 0.6 m), highly colored systems (M =
190 PCU) with high DOC (M = 35.7 mg L"1) and very
low sulfate concentrations (M = 4.7 //eq L"1).
The lakes in the Northcentral Peninsula were rela-
tively deep (M = 3.6 m), with circumneutral pH (M
= 6.4), moderate color (M = 42.4 PCU), moderate
ANC (M = 80.7 |xeq L~1), and DOC concentrations
(M = 9.0 mg LT1), high calcium (M = 225.1 n.eq L"1),
high sulfate concentrations (M = 191.4 n-eq L"1),
and very low silica concentrations (M = 0.37 jxeq
L~1). The concentrations of sodium and sulfate in
lakes that intercept deep groundwater flow are
likely related to karst aquifers (Heath and Conover
1981). However, the influence of atmospheric con-
tributions to regional aquatic chemistry has been
proposed for Northcentral Peninsula seepage lakes
that are not associated with karst groundwater in-
puts (Hendry and Brezonik 1984).
The lakes of the Southcentral Peninsula had higher
sulfate (M = 454.9 jj-eq L"1) and calcium concentra-
tons (M = 308.2 n,eq L~1) and lower concentrations
of DOC (M = 3.8 mg L~') than those in the North-
central Peninsula. Canfield (1983) and Canfield et al.
(1984) also found the greatest variability of pH and
ANC in central Florida because of diverse geology
and physiography.
Results of the current study show lakes along the
Gulf Coast of Florida had high pH (M = 6.9) and
high ANC (M = 414.2 (xeq L~1) and calcium
(M = 463.2 neq L~1) concentrations. Sulfate and
DOC concentrations were intermediate (M = 54.1
fjieq L~1 and 9.8 mg L"1, respectively) with respect
to all lakes in the Florida subregion.
117
-------
Figure 6-1. Classes of ANC (jueq L"1) in five selected subpopulations of lakes within Florida (3B), Eastern Lake Survey -Phase I.
Southcentral
Peninsula
118
-------
Section 7
Summary Observations8
7.1 Objectives
The primary objectives of the Eastern Lake Survey
were to:
1. determine the percentage (by number and
area) and location of lakes in potentially sensi-
tive regions of the U.S. that are acidic;
2. determine the percentage (by number and
area) and location of lakes in potentially sensi-
tive regions of the U.S. that have low acid neu-
tralizing capacity; and
3. provide the data base for selecting regionally
characteristic lakes for further study in
Phases II (temporal variability and biological
resources) and III (long-term monitoring) of
the project.
The summary observations presented in Sections
7.2 and 7.3 address the first two objectives. The
remaining summary observations address the third
objective of the Survey. These observations lead to
hypotheses that can be tested in subsequent
phases of the National Surface Water Survey and/
or the Aquatic Effects Research Program.
7.2 Extent and Location of Acidic and
Low pH Lakes
The subregions in the Eastern U.S. that contain the
largest proportion of acidic (ANC ^0 p-eq L~1) and
low pH (<5.0) lakes are the Adirondacks (1A), the
Upper Peninsula of Michigan (2B) and Florida (3B).
7.2.1 Acidic Lakes
• Within the Northeast (Region 1), the Adiron-
dacks (1A) had the largest estimated number
(138) and percentage (11%) of lakes with ANC
<0 n,eq L~1, followed by Southern New England
(1D; 5%), and the Poconos/Catskills (1B; 5%)
(Section 4.5.1.2). Maine (1E) had the lowest per-
centage of acidic lakes «1 %). Most acidic lakes in
the Adirondacks (1A) occurred in the western
portion of the subregion (Section 6.1.1).
• In the Upper Midwest (Region 2), 10 percent of
the lakes in the Upper Peninsula of Michigan
(2B) had ANC <0 (xeq L~1, and three percent in
Northcentral Wisconsin (2C) were acidic (Sec-
tion 4.5.1.3). In Northeastern Minnesota (2A)
and the Upper Great Lakes Area (2D) no acidic
lakes were sampled.
• In the Southeast (Region 3), no acidic lakes
were sampled in the Southern Blue Ridge (3A)
(Section 4.5.1.4). In contrast, an estimated 22
percent of the lakes in Florida (3B) had ANC <0
8The numbers and percentages of lakes cited here are population esti-
mates/
Acidic lakes in the Northeast had higher con-
centrations of sulfate, calcium, and extractable
aluminum than did acidic lakes in the Upper
Midwest and Southeast (Section 5.4).
7.2.2 Low pH Lakes
The estimated number of lakes and lake area with
low pH (pH ^5.0) also varied substantially among
and within regions.
• Within the Northeast, the Adirondacks (1 A) had
the largest estimated number (128) and per-
centage (10%) of lakes with pH <5.0. Subregion
1D (Southern New England) contained the sec-
ond highest estimated number (66) and per-
centage (5%) and the largest area (2295 ha, 6%)
of low pH lakes (Section 4.5.2.2). Maine (1 E) had
the fewest lakes (8, <1%) and least area (95 ha)
with pH <5.0.
• In the Upper Midwest, no lakes with pH <5.0
were observed in Northeastern Minnesota (2A)
or the Upper Great Lakes Area (2D) (Section
4.5.2.3). The Upper Peninsula of Michigan (2B)
was estimated to contain 99 lakes with pH <5.0,
representing nearly the same proportion as in
the Adirondacks (9% and 10%, respectively).
• In the Southeast, no lakes with pH <5.0 were
sampled in the Southern Blue Ridge (3A) (Sec-
tion 4.5.2.4). Florida (3B) had the highest esti-
mated number and percentage of lakes (259,
12%) and the largest estimated lake area with
pH <5.0.
119
-------
7.3 Extent and Location of Low ANC
Lakes
As observed with the estimates of low pH lakes, the
estimated number of lakes with low ANC varied
among and within regions.
• Within the Northeast, the Adirondacks (1 A) con-
tained the highest percentages of lakes with ANC
<50 /ueq L'1 and <200 /ueq L'1 (36% and 70%,
respectively) (Section 4.5.1.2). Central New Eng-
land (1C) and Maine (1E) contained the next
highest percentages of lakes among all eleven
subregions with ANC <200 /ueq L'1
subregions
67%, respectively).
/ueq L'1 (68% and
• Northcentral Wisconsin (2C) contained the
highest percentage (41 %) of lakes with ANC <50
peq L"1 among all eleven subregions (Section
4.5.1.3). Northeastern Minnesota (2A) and
Northcentral Wisconsin (2C) contained the high-
est percentage of lakes in the Upper Midwest
with ANC <200 ,ueq L"1 (57%). Although the
Upper Great Lakes Area (2D) contained the lowest
percentages in the Upper Midwest of lakes with
ANC <200 /ueq L"1, it contained the largest
number of lakes among all eleven subregions in
this category (1411).
• The Southern Blue Ridge (3A) contained the
lowest percentage (1%) and number (4) of lakes
with ANC <50 /aeq L~1 and the lowest number of
lakes with ANC <200 /ueq L"1 among all eleven
subregions (Section 4.5.1.4). Florida (3B) con-
tained the highest number of lakes among all
subregions with ANC <50 /ueq L"1, and the
second highest number of lakes with ANC <200
//eq I"1.
7.4 Chemical Characterization
7.4.7 Sulfate
Sulfate concentrations in lakes were greatest in
Florida and the southern portions of the Northeast.
No linear relationship between lakewater sulfate
and pH or ANC was evident in any region. High
concentrations of sulfate were found at low and
high pH values.
• Sulfate concentrations were relatively high in
the Northeast (median concentration (M)
= 115.4 fjueq L"1) (Section 4.6.1). Within the
Northeast, sulfate concentrations were highest
in the Poconos/Catskills (1 B; M = 1 59.3 /ueq L"1)
and Southern New England (1D; M = 141.1 jxeq
L~1). The lowest sulfate values were observed
in Maine (1E; M = 74.6 n-eq L~1).
varied among subregions within the Upper
Midwest, ranging from 50.1 (jieq L"1 in the
Upper Great Lakes Area (2D) to 77.7 jjieq L~1 in
the Upper Peninsula of Michigan (2B).
• In the Southeast, the Southern Blue Ridge (3A)
contained few lakes with high sulfate (22 or 8%
with S04"2 >150 /ueq L"1) (Section 4.6.1). This
subregion also had the lowest median sulfate
concentration, 31.8 jxeq L~1. Florida (3B) con-
tained the largest number of lakes with high
sulfate concentrations (846 or 40% with S04~2
a150 fjueq L~1). Subregion 3B also had the most
variable sulfate concentrations of any subre-
gion (Qd = 244 (xeq L~1).
7.4.2 Calcium
Calcium concentrations were lowest in the Upper
Midwest and Florida lakes.
• Within the Northeast, Southern New England
(1D) had the highest percentage and number of
lakes with calcium concentrations <50 ixeq L~1
(10%; 133) (Section 4.6.2). The Adirondacks
(1A) contained the second highest percentage
and number (8%; 108) of low calcium lakes
(s50 ixeq L"1).
The median sulfate concentration in the Upper
Midwest was half that of the Northeast (Section
4.6.1). Median sulfate concentrations also
120
• Northcentral Wisconsin (1C) contained the
highest percentage (22%) and second highest
number (34) of low calcium lakes among all
subregions (Section 4.6.2). The Upper Penin-
sula of Michigan (2B) contained the second
highest percentage (16%) of low calcium lakes
and the Upper Great Lakes Area (2D) contained
the second highest number (256) of low cal-
cium lakes in the Upper Midwest.
• In the Southeast, 12 percent of the lakes in the
Southern Blue Ridge (3A) had low concentra-
tions of calcium, whereas in Florida (3B), 19 per-
cent of the lakes were in this group (Section
4.6.2). Florida (3B) contained the highest number
(402) of low calcium lakes among all subre-
gions.
7.4.3 Extractable Aluminum
Extractable aluminum concentrations were higher
in lakes with lower pH values, and higher in the
Northeast than in other regions.
• The largest estimated number of clearwater
lakes having extractable aluminum concentra-
tions >150 |xg L"1 occurred in the Adirondacks
(1A; 82 lakes or 10%) (Section 4.6.3). Few lakes
in the Poconos/Catskills (1B; 3 lakes or <1%)
and Southern New England (1D; 7 lakes or 1%)
had extractable aluminum a150 (xg L~1. No
clearwater lakes sampled in Maine (1E) had ex-
tractable aluminum concentrations s50 jig L"1.
-------
• Extractable aluminum concentrations in clear-
water lakes were lower in the Upper Midwest
(Q4 = 8.5 (xg L~1) than in the Northeast (Q4
= 11.6 |xg L~1) (Section 4.6.3). Extractable alu-
minum was lowest in clearwater lakes in North-
eastern Minnesota (2A; 04 = 3.0 ng L~1), and
highest in clearwater lakes in the Upper Penin-
sula of Michigan (2B; Q4= 11.9 fig L~1).
• Extractable aluminum concentrations in clear-
water lakes were low in the Southern Blue
Ridge (3A; Q4 = 2.5 ng L~1) (Section 4.6.3). In
Florida (3B), clearwater lakes had lower ex-
tractable aluminum concentrations (Q4=18.6
(xg L~1) than did clearwater lakes in the Adiron-
dack subregion (1A; Q4 = 29.4 fig L~1) (Sec-
tion 5.4).
• In each region extractable aluminum concen-
trations were higher at lower pH values (Sec-
tion 5.2.3.2). The Northeast had the greatest in-
crease in extractable aluminum with
decreasing pH and Florida the least increase at
low pH values.
7.4.4 Dissolved Organic Carbon
Dissolved organic carbon (DOC) concentrations did
not correlate with the distribution of acidic or low
ANC lakes.
• In the Northeast, as in other regions, 80 percent
of acidic lakes contained concentrations of DOC
<5 mg L~1 (Section 5.4). A positive relationship
existed between pH and DOC. Those lakes with
highest DOC concentrations were drainage
lakes with short hydraulic residence times and
high ANC (Section 5.3.2).
• In the Upper Midwest, most acidic lakes, espe-
cially those in the Upper Peninsula of Michigan
(2B) and Northcentral Wisconsin (2C), were
clearwater, low DOC, seepage lakes (Section
6.2). Lakes in Northeastern Minnestoa (2A) had
the highest concentrations of DOC in the Upper
Midwest and no acidic lakes were sampled in
this subregion (Sections 4.5.1 and 4.6.4).
• In the Southeast, only the lakes within the Oke-
fenokee Swamp exhibited a strong association
between low pH and high DOC. No apparent
relationship between pH and DOC was evident
in Florida (3B) lakes (Section 6.3.2).
7.4.5 Major Cations and Anions
The anions were most useful in characterizing dif-
ferences in the relative importance of major ions
among regions and subregions.
• In the Northeast, sulfate was the predominant
anion at the 20th percentile in three of the sub-
regions (Adi rondacks, 1 A; Poconos/Catskills, 1B;
and Central New England, 1C) (Section 5.2.5.2).
Sulfate was also the dominant anion at the
median value in the Adirondacks (1 A).
• In Maine (1E), bicarbonate ion concentrations
exceeded sulfate at both the 20th percentile and
the median (Section 5.2.5.2).
• Chloride was the dominant anion in Southern
New England (1D) at both the 20th percentile
and median values estimated for the popula-
tion (Section 5.2.5.2).
• Bicarbonate was the dominant anion at the 20th
percentile and median values in the Upper Mid-
west, with the exception of the Upper Peninsula
of Michigan (2B) and Northcentral Wisconsin
(2C), where sulfate was dominant at the 20th
percentile (Section 5.2.5.2).
• The ionic composition of lakes in. Florida (3B)
was similar to that of lakes in Southern New
England (1D) in that sodium was the dominant
cation and chloride the dominant anion at the
20th percentile (Section 5.2.5.2). Total ionic
concentration of many Florida lakes was high.
• Organic anions, as indicated by anion deficit,
were not the dominant anions in any subregion
at either the 20th or 50th percentiles (Section
5.2.6.2). Concentrations of organic anions were
especially low in the Northeast.
7.5 Future Studies
The results of the Survey presented in this report
are largely descriptive. However, the statistical de-
sign of the Survey makes it possible to test acidifi-
cation hypotheses and to examine the results with
respect to a population of lakes. Future aspects of
the National Lake Survey combined with other data
bases can address, in part, the following questions.
• Sulfate concentrations in lakes across the
Northeast and the Upper Midwest show an ap-
parently strong relationship with the general
patterns of sulfate deposition as measured by
the National Trends Network. What is the na-
ture of the relationship between lake chemistry
and atmospheric deposition of sulfate?
• The majority of acidic lakes in all three regions
contained relatively low concentrations of or-
ganic acids. How important are the contribu-
tions of organic acids in explaining the occur-
rence of acidic lakes?
• Some portions of the coastal areas of the North-
east contained moderate numbers of acidic
lakes. To what degree can the acidity of these
coastal lakes be attributed to a neutral salt ef-
fect from sea spray deposition ?
121
-------
The estimated hydraulic residence times for
clearwater lakes were approximately 3 times
greater than for darkwater lakes. Residence time
was inversely related to DOC. Does an apparent
difference in hydrology between clearwater,
acidic lakes and darkwater, higher ANC lakes
indicate that acidic lakes generally are not derived
from darkwater lakes?
Florida (3B) contained the largest proportion of
acidic lakes and their chemistry differed consid-
erably in many respects from lakes in the North-
east, Upper Midwest and Southern Blue Ridge
(3A). To what degree are the acidic lakes in Flor-
ida affected by acidic deposition, and are other
factors more important in explaining the occur-
rence of acidic lakes in Florida!
122
-------
Section 8
References
1. Aimer, B., W. Dickson, C. Ekstrom, and E.
Hornstrom. Sulfur Pollution and the Aquatic
Ecosystem. In: Sulfur in the Environment:
Part II, Ecological Impacts, J. O. Nriagu, ed.
John Wiley & Sons, New York, New York,
1978. pp. 271-311.
2. Altshuller, A. P., and G. A. MacBean. Second
Report of the U.S.-Canada Research Consulta-
tion Group on Long Range Transport of Air
Pollutants. November 1980. (Available from
Envir. Can., Ottawa.)
3. Altshuller, A. P., and R. A. Linthurst (eds.). The
Acidic Deposition Phenomenon and Its Ef-
fects. Critical Assessment Review Papers,
Vol. II: Effects Sciences. EPA-600/8-83-016 BF,
U.S. Environmental Protection Agency, Wash-
ington, D.C., 1984.
4. American Society for Testing and Materials
(ASTM). Annual Book of ASTM Standards.
Vol. 11.01, Standard Test Methods for Anions
in Water by Ion Chromatography. ASTM,
Philadelphia, Pennsylvania, 1984. pp. 4327-84.
5. Anonymous. Acid Sensitivity Survey of Lakes
in Ontario. Ontario Ministry of the Environ-
ment. APIOS 002/81, 1981.
6. Bailey, J. H. Proposed Trophic Classification
of the Great Ponds of Maine. Maine Depart-
ment of Environmental Protection, Augusta,
Maine, 1975.
7. Baker, J. P. Chapter 5. Fish. In: The Acidic Dep-
osition Phenomenon and Its Effects. Critical
Assessment Review Papers, Vol. II: Effects
Sciences, A. P. Altshuller and R. A. Linthurst,
eds. EPA-600/8-83-016 BF, U.S. Environmental
Protection Agency, Washington, D.C., 1984.
pp. 5-74 to 5-129.
8. Baker, J. P., and C. L. Schofield. Aluminum
Toxicity to Fish in Acidic Waters. Water, Air,
and Soil Poll., 18:289-309, 1982.
9. Baker, L. A. Mineral and Nutrient Cycles and
Their Effect on the Proton Balance of a Soft-
water, Acidic Lake. Ph.D. Thesis, University of
Florida, Gainesville, Florida, 1984. 151 pp.
10. Baker, L. A., P. L. Brezonik, and C. R. Kratzer.
Nutrient Loading-Trophic State Relationships
in Florida Lakes. Water Resoures Research
Center Publication Number 56. University of
Florida, Gainesville, Florida. 1981.
11. Baker, L. A., P. L. Brezonik, E. S. Edgerton, and
W. O. Ogburn. Sediment Acid Neutralization
in Softwater Lakes. Water, Air, and Soil Poll.,
25:215-230, 1985.
12. Barnes, R. B. The Determination of Specific
Forms of Aluminum in Natural Water. Chem.
Geol., 15:177-191, 1975.
13. Barr, K. Estimation of Groundwater and Lake
Chemistry Parameters and Their Interrelation-
ships in Glaciated Terranes. M.S. Thesis, Uni-
versity of Minnesota, 1978. 293 pp.
14. Bayly, I. A. E. Chemical and Biological Studies
on Some Acidic Lakes of East Australian
Sandy Coastal Lowlands. Aust. J. Mar.
Freshw. Res., 15:56-72, 1964.
15. Beck, K. C., J. H. Reuter, and E. M. Perdue.
Organic and Inorganic Geochemistry of Some
Coastal Plain Rivers of the Southeastern
United States. Geochim. Cosmochim. Acta,
38:341-364, 1974.
16. Best, M. D., L. W. Creelman, S. K. Drouse, and
D. J. Chaloud. National Surface Water Survey,
Eastern Lake Survey-Phase I, Quality Assur-
rance Report. EPA-600/4-86-011, U.S. Environ-
mental Protection Agency, Las Vegas, Ne-
vada, 1986.
17. Black, A. P., and R. F. Christman. Characteris-
tics of Colored Surface Waters. J.A.W.W.A.,
55:753-770, 1963.
18. Blank, G. B., P. J. Dillon, and J. D. Gregory.
Chapter 4. Alkalinity as An Indicator of Sensi-
tivity. In: The Acidic Deposition Phenomenon
and Its Effects. Critical Assessment Review
Papers, Vol. II: Effects Sciences, A. P. Alt-
shuller and R. A. Linthurst, eds. EPA-600/8-83-
016 BF, 1984. pp. 4-24 to 4-28.
19. Bradt, P. T., and M. B. Berg. Preliminary Sur-
vey of Pocono Mountain Lakes to Determine
123
-------
Sensitivity to Acid Deposition. Proc. Pa. Acad.
Sci., 57:190-194, 1983.
20. Bradt, P. T., M. B. Berg, D. S. Barrasso, and
J. L Dudley. The Biological and Chemical Im-
pact of Acid Precipitation on Pocono Moun-
tain Lakes. Unpublished Report. Department
of Biology, Lehigh University, Bethlehem,
Pennsylvania. 1984.
21. Brezonik, P. L, C. D. Hendry, Jr., E. S. Edger-
ton, R. L. Schulze, and T. L. Crisman. Acidity,
Nutrients, and Minerals in Atmospheric Pre-
cipitation Over Florida: Deposition Patterns,
Mechanisms and Ecological Effects. EPA-600/
53-83-004, U.S. Environmental Protection
Agency, Corvallis, Oregon, June 1983.
22. Brooks, J. L., and E. S. Deevey, Jr. Chapter 4.
New England. In: Limnology in North Amer-
ica, D. G. Frey, ed. University of Wisconsin
Press, Madison, Wisconsin, 1963. pp. 117-162.
23. Broughton, W. A. The Geology, Groundwater
and Lake Basin Seal of the Region South of the
Muskellunge Moraine, Vilas County, Wiscon-
sin. Trans. Wis. Acad. Sci., Arts, and Lettr.,
33:5-20, 1941.
24. Brugam, R. B. Chemistry of Lake Water and
Groundwater in Areas of Contrasting Glacial
Drift. Hydrobiologia, 80:47-62, 1981.
25. Busby, M. W. Annual Runoff in the Contermin-
ous United States. Atlas HA-212 Hydrologic
Investigations. United States Geological Sur-
vey. U.S. Department of the Interior, 1966.
26. Butler, J. N. Carbon Dioxide Equilibria and
Their Applications. Addison-Wesley Publish-
ing, Reading, Massachusetts, 1982. 259 pp.
27. Canfield, D. E., Jr. Chemical and Trophic State
Characteristics of Florida Lakes: Relation to
Regional Geology. Institute of Food and Agri-
cultural Sciences, University of Florida,
Gainesville, Florida, 1981.
28. Canfield, D. E., Jr. Sensitivity of Florida Lakes
to Acidic Precipitation. Water Rsources Re-
search, 19:833-839, 1983.
29. Canfield, D. E., Jr., S. B. Linda, and L M. Hodg-
son. Relations Between Color and Some Lim-
nological Characteristics of Florida Lakes.
Water Resources Bulletin, 20:323-329, 1984.
30. Chen, C. W., S. A. Gherini, N. E. Peters, P. S.
Murdock, R. M. Newton, and R. A. Goldstein.
Hydrologic Analysis of Acidic and Alkaline
Lakes. Water Resources Research, 20:1875-
1882, 1984.
31. Church, M. R., and J. N. Galloway. Application
of Henriksen's "Acidification Indicator" and
"Predictor Nomograph" to Two Adirondack
Lakes. Water, Air, and Soil Poll., 22:111-120,
1984.
32. Clarkson, B. Vermont Acid Precipitation Pro-
gram Winter Lake Surveys, 1980-1982. Ver-
mont Department of Water Resources and En-
vironmental Engineering, 1982. 28 pp.
33. Colquhoun, J., W. Kretser, and M. Pfeiffer.
Acidity Status Update of Lakes and Streams in
New York State. New York State Department
of Environmental Conservation, Division of
Fish and Wildlife, April 1984.
34. Cook, R. B., C. A. Kelley, R. G. Kreis, Jr., and
J. C. Kingston. Chemical Limnology of Lakes
in the Northern Great Lakes States. Unpub-
lished Report. Department of Geoscience, In-
diana University of Pennsylvania, Indiana,
Pennyslvania, 1985.
35. Cook, R. B., and D. W. Schindler. The Biogeo-
chemistry of Sulfur in an Experimentally Acid-
ified Lake. Ecol. Bull. (Stockholm), 35:115-127,
1983.
36. Cooper, G. P. A. Biological Survey of Lakes
and Ponds of the Central Coastal Area of
Maine. Maine Department of Inland Fish and
Game. Fish Survey Report 5:1-185, 1942.
37. Cronan, C. S., and C. L. Schofield. Aluminum
Leaching Response to Acid Precipitation: Ef-
fects on High-elevation Watersheds in the
Northeast. Science, 204:304-306, 1979.
38. Davis, R. B., M. O. Smith, J. H. Bailey, and
S. A. Norton. Acidification of Maine (USA)
Lakes by Acidic Precipitation. Verh. Internal.
Verein. Limnol., 10:532-537, 1978a.
39. Davis, R. B., J. H. Bailey, M. Scott, G. Hunt, and
S. A. Norton. Descriptive and Comparative
Studies of Maine Lakes. University of Maine
Life Sciences and Agricultural Experiment
Station Technical Bulletin No. 88, Orono,
Maine, 1978b. 337 pp.
40. Deevey, E. S. Limnological Studies in Con-
necticut: A Contribution to Regional Limnol-
ogy. Amer. J. Sci., 238:717-741, 1940.
41. Dickson, W. The Acidification of Swedish
Lakes. Inst. Freshwater Res. Drottingholm.
Rep., 54:8-20, 1975.
42. Dickson, W. Some Effects of the Acidification
of Swedish Lakes. Verh. Internet. Verein.
Limnol., 20:851-856, 1978.
124
-------
43. Drablos, D., and A. Tollan, eds. Ecological Im-
pact of Acid Precipitation. Proceedings of an
International Conference, Sandefjord, Nor-
way, SNSF Project, March 11-14, 1980.
44. Drever, J. I. The Geochemistry of Natural
Waters. Prentice Hall, Inc. Englewood Cliffs,
New Jersey, 1982. p. 47.
45. DriscolS, C. T. A Procedure for the Fractiona-
tion of Aqueous Aluminum in Dilute Acidic
Waters. Internat. J. Environ. Anal. Chem.,
16:267-283, 1984.
46. Driscoll, C. T., J. P. Baker, J. J. Bisogni, and
C. L Schofield. Effect of Aluminum Speciation
on Fish in Dilute Acidified Waters. Nature,
284:161-164, 1980.
47. Driscoli, C. T., J. P. Baker, J. J. Bisogni, and
C. L. Schofield. Aluminum Speciation and
Equilibria in Dilute Acidic Surface Waters of
the Adirondack Region of New York State. In:
Geological Aspects of Acid Deposition, 0. P.
Bricker, ed. Butterworth Publishers, Boston,
Massachusetts, 1983. pp. 55-75.
48. Drouse, S. K., D. C. J. Hillman, L. W. Creelman,
J. F. Potter, and S. J. Simon. National Surface
Water Survey, Eastern Lake Survey-Phase I,
Quality Assurance Plan. EPA-600/4-86-008,
U.S. Environmental Protection Agency, Las
Vegas, Nevada, 1986.
49. Eilers, J. M., G. E. Glass, K. E. Webster, and
J. A. Rogalla. Hydrologic Control of Lake Sus-
ceptibility to Acidification. Can. J. Fish. Aq.
Sci., 40:1896-1904, 1983.
50. Eilers, J. M., D. J. Blick, Jr., and M. D. DeHaan.
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, 1986.
51. Environmental Data Service. Climatography
of United States #81 (by state) Monthly Nor-
mals of Temperature, Precipitation, and Heat-
ing and Cooling Degree Days. 1951-1980.
National Oceanic and Atmospheric Adminis-
tration, U.S. Department of Commerce, 1983.
52. Environmental Science and Engineering, Inc.
Florida Acid Deposition Study. Phase II Report
Addendum: Acid Deposition Monitoring Pro-
gram. Data Analysis. ESE No. 81-631-600.
Gainesville, Florida, November 1983.
53. Eriksson, E. The Yearly Circulation of Chloride
and Sulfur in Nature; Meterological, Geo-
chemical, and Pedological Implications. Tel-
lus, 12:63-109, 1960.
54. Eshleman, K. N., and H. F. Hemond. The Role
of Organic Acids in the Acid Base Status of
Surface Waters at Bickford Watershed, Mas-
sachusetts. Water Resources Research,
21:1503-1510, 1985.
55. Field, R., E. J. Struzeski, Jr., H. E. Masters, and
A. N. Tafuri. Water Pollution and Associated
Effects from Street Salting. EPA-R2-73-257,
U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1973.
56. Galloway, J. N. Chapter 4. Alkalinity as An In-
dicator of Sensitivity, Relative Importance of
HN03vs. H2S04and Long-Term Acidification.
In: The Acidic Deposition Phenomenom and
Its Effects. Critical Assessment Review
Papers, Vol. II: Effects Sciences, A. P. Alt-
shuller and R. A. Linthurst, eds. EPA-600/8-83-
016 BF, U.S. Environmental Protection
Agency, Washington, D.C., 1984. pp. 4-6 to
4-7,4-24 to 4-28,4-39 to 4-45, and 4-48 to 4-53.
57. Gilliom, R. J., R. M. Hirsch, and E. J. Gilroy.
Effect of Censoring Trace-level Water Quality
Data on Trend Detection Capability. Env. Sci.
Technol., 18(7): 530-535, 1984.
58. Gjessing, E. T. Physical and Chemical Charac-
teristics of Aquatic Humus. Ann Arbor Science
Publishers, Ann Arbor, Michigan, 1976.
120pp.
59. Glass, G. E., and O. L. Loucks. Implications of
a Gradient in Acid and Ion Deposition Across
the Northern Great Lakes States. Environ. Sci.
Technol., 20(1):35-43, 1986.
60. Gorham, E. On Some Factors Affecting the
Chemical Composition of Swedish Fresh
Waters. Geochim. Cosmochim. Acta, 7:129-
150, 1955.
61. Gorham, E., S. E. Bayley, and D. W. Schindler.
Ecological Effects of Acid Deposition Upon
Peatlands: A Neglected Field of "Acid-rain"
Research. Can. J. Fish. Aq. Sci., 41:1256-1268,
1984.
62. Gorham, E., W. E. Dean, and J. E. Sanger. The
Chemical Composition of Lakes in the North-
central United States. Limnol. Oceanogr.,
28:287-301, 1983.
63. Gorham, E., S. J. Eisenreich, J. Ford, and M. V.
Santelmann. The Chemistry of Bog Waters.
In: Chemical Processes in Lakes, W. Strumm,
ed. John Wiley & Sons, New York, New York,
1985. pp. 339-363.
64. Hadley, D. W., and J. H. Pelham. Glacial De-
posits of Wisconsin: Sand and Gravel Re-
source Potential, 1:500,000. Map 10. Wiscon-
125
-------
sin Geological Natural History Survey,
Madison, Wisconsin, 1976.
65. Haines, T. A., and J. J. Akielaszek. A Regional
Survey of the Chemistry of Headwater Lakes
and Streams in New England: Vulnerability to
Acidification. In: Air Pollution and Acid Rain.
Report No. 15, U.S. Department of the Interior,
Fish and Wildlife Service, FWS/OBS-80/40/15,
1983, 141 pp.
66. Hanes, R. E., L W. Zelazny, and R. E. Blaser.
Effects of Deicing Salts on Water Quality and
Biota, Literature and Recommended Re-
search. National Cooperative Highway Re-
search Program Report 91, 1970.
67. Heath, R. C., and C. S. Conover. Hydrologic
Almanac of Florida. U.S. Geological Survey
Open File Report 81-1107, 1981. 239 pp.
68. Hem, J. D. Study and Interpretation of the
Chemical Characteristics of Natural Water.
2nd Edition. Washington, D.C.: Geological
Survey Supply Paper 1473, 1970.
69. Hendry, C. D., and P. L Brezonik. Chemical
Composition of Softwater Florida Lakes and
Their Sensitivity to Acid Precipitation. Water
Resources Bulletin, 20:75-86, 1984.
70. Hendrey, G. R., J. N. Galloway, and C. L.
Schofield. Temporal and Spatial Trends in the
Chemistry of Acidified Lakes Under Ice Cover.
In: Ecological Impact of Acid Precipitation,
D. Drablos and A. Tollan, eds.. Proceedings of
an International Conference, Sandefjord, Nor-
way, SNSF Project, 1980a. pp. 266-267.
71. Hendrey, G. R., J. N. Galloway, S. A. Norton,
C. L Schofield, P. W. Shaffer, and D. A. Burns.
Geological and Hydrochemical Sensitivity of
the Eastern U.S. to Acid Precipitation. EPA-
600/3-80-024, U.S. Environmental Protection
Agency, Corvallis, Oregon, 1980b.
72. Henriksen, A. A Simple Approach for Identify-
ing and Measuring Acidification of Fresh-
water. Nature, 278:542-545, 1979.
73. Henriksen, A. Acidification of Freshwaters-A
Large Scale Titration. In: Ecological Impact of
Acid Precipitation, D. Drablos and A. Tollan,
eds.. Proceedings of an International Confer-
ence, SNSF-project, Sandefjord, Norway,
1980. pp. 68-74.
74. Henriksen, A. Changes in Base Cation Concen-
trations Due to Freshwater Acidification. Acid
Rain Research Report #1, NIVA, Oslo, 1982.
50pp.
75. Hillman, D. C. J., J. F. Potter, and S. J. Simon.
National Surface Water Survey, Eastern Lake
Survey-Phase I, Analytical Methods Manual.
EPA-600/4-86-009, U.S. Environmental Protec-
tion Agency, Las Vegas, Nevada, 1986.
76. Holland, H. D. The Chemistry of the Atmo-
sphere and Oceans. John Wiley and Sons,
New York, New York, 1978. pp. 103-107.
77. Hutchinson, G. E. A Treatise on Limnology.
Vol. I, Geography, Physics and Chemistry.
John Wiley & Sons, New York, New York,
1957. 1015 pp.
78. ll'in, N. P., and D. S. Orlov. Photochemical De-
struction of Humic Acids. Soviet Soil Sci., 75-
83, 1973.
79. Jackson, D. D. The Normal Distribution of
Chlorine in the Natural Waters of New York
and New England. U.S. Geological Survey.
Water Supply Paper No. 144, 1905. 31 pp.
80. Johnson, D. W., and J. O. Reuss. Soil-
mediated Effects of Atmospherically De-
posited Sulphur and Nitrogen. Phil. Trans. R.
Soc. London, England, B 305:383-392, 1984.
81. Juday, C., E. A. Birge, and V. W. Meloche. The
Carbon Dioxide and Hydrogen Ion Content of
the Lake Waters of Northeastern Wisconsin.
Trans. Wise. Acad. Sci., Arts, and Lettr., 29:1-
82, 1935.
82. Junge, C. E., and R. T. Werby. The Concentra-
tion of Chloride, Sodium, Potassium, Calcium,
and Sulfate in Rain Water Over the United
States. J. Meterol., 15:417-425, 1958.
83. Kanciruk, P., R. J. Olson, and R. A. McCord.
Data Quality Assurance in a Research Data
Base: The 1984 National Surface Water Sur-
vey. In: Research Data Management in the
Ecological Sciences. Belle W. Baruch Library
and Marine Science No. 16, W. K. Michener,
ed.. University of South Carolina Press, Co-
lumbia, South Carolina, 1986a. pp. 193-207.
84. Kanciruk, P., R. A. McCord, L. A. Hook, and M.
J. Gentry. National Surface Water Survey,
Eastern Lake Survey - Phase I, Data Base Dic-
tionary. Oak Ridge National Laboratory, Tech-
nical Manual, Oak Ridge National Laboratory,
Oak Ridge, Tennessee, (in prep.), 1986b.
85. Kramer, J. R. Modified Gran Analysis for Acid
and Base Titrations. Environmental Geochem-
istry Report No. 1984-2. McMaster University,
Hamilton, Ontario, Canada, 1984.
86. Kramer, J. R., and A. Tessier. Acidification of
Aquatic Systems: A Critique of Chemical Ap-
proaches. Env. Sci. Technol., 16:606A-615A,
1982.
726
-------
87. Kroopnick, P. The S04:CI Ratio in Oceanic
Rainwater. Pacific Sci., 31:91-106, 1977.
88. Krug, E. C., P. J. Isaacson, and C. R. Frink.
Appraisal of Some Current Hypotheses De-
scribing Acidification of Watersheds. Pre-
sented at the 77th Annual Meeting of the Air
Pollution Control Association, San Francisco,
California, June 24-29, 1984.
89. Krug, E. C., and C. R. Frink. Acid Rain on Acid
Soil: A New Perspective. Science, 221:520-
525, 1983.
90. Lebowitz, L G., and R. G. de Pena. Chloride
and Sodium Content in Northeastern United
States Precipitation. J. Geophys. Res.,
90:8149-8154, 1985.
91. Lamar, W. L, and D. F. Goerlitz. Organic Acids
in Naturally Colored Surface Waters. Geologi-
cal Survey Water-Supply Paper 1817-A, 1966.
17pp.
92. Liaw, W. K. Chemical Sensitivity of Northern
Saskatchewan Waters to Acidification. In: Pro-
ceedings on Emissions in Alberta and the Eco-
logical Effects, Edmondton, Alberta, Canada,
1982. pp. 485-510.
93. Lillie, R. A., and J. W. Mason. Limnological
Characteristics of Wisconsin Lakes. Technical
Bulletin No. 138, Wisconsin Department of
Natural Resources, 1983. 116 pp.
94. Lin, J. C., and J. L. Schnoor. An Acid Precipita-
tion Model for Seepage Lakes. J. Env. Engr.
Div. ASCE, (in press), 1986.
95. Livingstone, D. A. Chemical Composition of
Rivers and Lakes. Chapter G. Data of Geo-
chemistry. 6th ed. Professional Papers U.S.
Geological Survey, 440-G, 1963. 64 pp.
96. Mairs, D. F. A Total Alkalinity Atlas for Maine
Lake Waters. Limnol. Oceanogr., 11:68-72,
1966.
97. Mairs, D. F. Surface Chloride Distribution in
Maine Lakes. Water Resources Research,
3:1090-1092, 1967.
98. Morris, F. A., D. V. Peck, M. B. Bonoff, and K.
J. Cabbie. National Surface Water Survey,
Eastern Lake Survey - Phase I, Field Opera-
tions Report. EPA-600/4-86-010, U.S. Environ-
mental Protection Agency, Las Vegas, Ne-
vada, 1986.
99. Moss, B. The Influence of Environmental Fac-
tors on the Distribution of Freshwater Algae:
An Experiment II. The Role of pH and the Car-
bon Dioxide-bicarbonate System. J. Ecol.,
61:157-177, 1973.
100. Moyle, J. B. Relationships Between the Chem-
istry of Minnesota Surface Waters and Wild-
life Management. J. Wild. Mgmt., 20:303-320,
1956.
101. Mudrey, M. G., Jr., B. A. Brown, and J. K.
Greenburg. Bedrock Geologic Map of Wiscon-
sin. University of Wisconsin - Extension and
Wisconsin Geological Natural History Survey,
Madison, Wisconsin, 1982.
102. National Academy of Sciences (NAS).
Atmospheric-biosphere Interaction: Toward a
Better Understanding of the Ecological Conse-
quences of Fossil Fuel Combustion. Academy
Press, Washington, D. C., 1981. 263 pp.
103. National Research Council Canada (NRCC).
Acidification in the Canadian Aquatic Environ-
ment. NRCC Publication No. 18475, 1981.
369 pp.
104. Norton, S. A. Distribution of Surface Waters
Sensitive to Acid Precipitation: A State Level
Atlas. NADP Technical Report #4, Department
of Geological Science, University of Maine,
Orono, Maine, 1982. 72 pp.
105. Norton, S. A., R. B. Davis, and D. F. Brakke.
Responses of Northern New England Lakes to
Atmospheric Inputs of Acid and Heavy Metals.
Completion Report Project A-048-ME, Land
and Water Resources Center, University of
Maine, Orono, Maine, 1981. 9^ pp.
106. O'Dell, J. W., J. D. Pfaff, M. E. Gales, and G. D.
McKee. Technical Additions to: Methods for
the Chemical Analysis of Water and Wastes.
Method 300.0. The Determination of Inorganic
Anions in Water by Ion Chromatography.
EPA-600/4-85-017, U.S. Environmental Protec-
tion Agency, Cincinnati, Ohio, 1984.
107. Ogden, J. G. Seasonal Mass Balance of Major
Ions in Three Small Watersheds in a Maritime
Environment. Water, Air, Soil Poll., 17:119-
130, 1982.
108. Oliver, B. G., E. M. Thurman, and R. L. Mal-
colm. The Contribution of Humic Substances
to the Acidity of Colored Natural Waters.
Geochim. Cosmochim. Acta, 47:2031-2035,
1983.
109. Omernik, J. M., and C. F. Powers. Total Alka-
linity of Surface Waters — A National Map.
Annals of the Association of American Geog-
raphers, 73:133-136, 1983.
110. Omernik, J. M. Total Alkalinity of Surface Wa-
ters: A Map of the Appalachian Region. U.S.
Environmental Protection Agency, Corvallis,
Oregon, (draft), 1985.
127
-------
111. Omernik, J. M., and G. E. Griffith. Total Alka-
linity of Surface Waters: A Map of the Upper
Midwest Region. EPA-600/D-85-043, U.S. En-
vironmental Protection Agency, Corvallis, Or-
egon, June 1985.
112. Omernik, J. M., and A. J. Kinney. Total Alkalin-
ity of Surface Waters: A Map of the New Eng-
land and New York Region. EPA-600/D-84-216,
U.S. Environmental Protection Agency, Cor-
vallis, Oregon, June 1985.
113. Omernik, J. M., A. J. Kinney, B. Abbruzzese, C.
B. Johnson, and B. J. Rosenbaum. The Na-
tional Surface Water Survey: A Regional Geo-
graphic Approach for Lake Selection. U.S. En-
vironmental Protection Agency, Corvallis,
Oregon, (draft), 1986.
114. Peters, N. E., R. A. Schroeder, and D. E. Trout-
man. Temporal Trends in the Acidity of Pre-
cipitation and Surface Waters of New York.
U.S. Geological Survey Water Supply Paper
2188, Washington, D.C., 1981.
115. Pfeiffer, M. H., and P. J. Festa. Acidity Status of
Lakes in the Adirondack Region of New York
in Relation to Fish Resources. New York State
Department of Environmental Conservation,
FW-P168(10/80), 1980.
116. Placke, J. Trophic State Evaluation of TVA
Reservoirs. TVA/ONR/WR-83/7. Tennessee
Valley Authority, Chattanooga, Tennessee,
1983.
117. Rapp, G., Jr., J. D. Albert, B. W. Liukkonen, J.
A. Use, O. L. Loucks, and G. E. Glass. Acid
Deposition and Watershed Characteristics in
Relation to Lake Chemistry in Northeastern
Minnesota. Env. Int., 11:425-440, 1985.
118. Roberson, C. E., and J. D. Hem. Solubility of
Aluminum in the Presence of Hydroxide, Fluo-
ride and Sulfate. U.S. Geological Survey
Water - Supply Paper 1827-C, 1969. 37 pp.
119. Rodhe, W. The Ionic Composition of Lake
Waters. Verh. Int. Verein. Limnol., 13:121-141,
1949.
120. Rosen, A. E., and P. Kanciruk. A Generic Data
Entry Quality Assurance Tool. Proceedings of
the Tenth Annual International SAS (Statisti-
cal Analysis System, Inc.) Users Group Con-
ference, Reno, Nevada, March 10-13,1985.
121. Rosenqvist, I. Th. Acid Precipitation and Other
Possible Sources for Acidification of Rivers
and Lakes. Sci. Total Environ., 10:271-272,
1978.
122. Ryder, R. A. Chemical Characteristics of On-
tario lakes Related to Glacial History. Trans.
Amer. Fish. Soc., 93:260-268, 1964.
123. SAS Institute. The Statistical Analysis System,
SAS Institute, Gary, North Carolina, 1982.
124. Schindler, D. W., and T. J. Ruszczynski. A Test
of Limnological Data from the Experimental
Lakes Area, Northwestern Ontario, for Evi-
dence of Acidification. Can. Tech. Rep. of Fish.
Aquatic Sci., No. 1147, Winnepeg, Manitoba,
Canada, 1983.
125. Schneider, J. C. Typology and Fisheries Po-
tential of Michigan Lakes. The Michigan Aca-
demician, 8(1):59-84, 1975.
126. Schnoor, J. L., W. Stumm, and J. J. Morgan.
Acid Rain: A Disruption of Hydrogeochemical
Processes. Naturwissenschaften, 70:216-223,
1983.
127. Schnoor, J. L., N. P. Nikolaidis, and G. E.
Glass. Lake Resources at Risk to Acidic Depo-
sition in the Upper Midwest. J. Wat. Poll.
Contr. Fed., 58:139-148, 1986.
128. Schofield, C. L. Acid Precipitation - Effects on
Fish. Ambio, 5:228-230, 1976.
129. Schofield, C. L. Historical Fisheries Changes in
the United States Related to Decreases in Sur-
face Water pH. In: Acid Rain/Fisheries, T. A.
Haines and R. E. Johnson, eds. American Fish-
eries Society, Bethesda, Maryland, 1982.
130. Shannon, E. E., and Brezonik, P. L. Limnologi-
cal Characteristics of North and Central Flor-
ida Lakes. Limnol. Oceanogr., 17:97-110,
1972.
131. Simpkins, W. W., M. C. McCartney, and D. M.
Mickelson. Glacial Geology of Forest County,
Wisconsin. Wisconsin Geological Natural His-
tory Survey Open File Report, Madison, Wis-
consin, 1978.
132. Skougstad, M. W., M. J. Fishman, L. C. Fried-
man, D. E. Erdman, and S. S. Duncan, eds.
Methods for Determination of Inorganic Sub-
stances in Water and Fluvial Sediments: Tech-
niques of Water-resources Investigations of
the United States Geological Survey, Book 5,
Chapter A1. U.S. Government Printing Office,
Washington, D.C., 1979.
133. Small, M. J., and M. C. Sutton. A Regional pH
- Alkalinity Relationship. Water Res., (in
press), 1986.
134. Stumm, W., and J. J. Morgan. Aquatic Chem-
istry: An Introduction Emphasizing Chemical
128
-------
Equilibria in Natural Waters. 2nd Edition. J Dhn
Wiley & Sons, New York, New York, 1981.
135. Thompson, M. E. Exchange of Marine Sodium
for Calcium During Chemical Weathering in
the Isle Aux Morts River Basin, Newfound-
land. Geochim. Cosmochim. Acta, 46:361-365,
1982.
136. Thornton, K. W., R. H. Kennedy, J. H. Carroll,
W. W. Walker, R. C. Gunkel, and S. Ashby.
Reservoir Sedimentation and Water Quality -
An Heuristic Model. In: Symposium on Sur-
face Water Impoundments, H. Stefan,
ASCE, Vol. I, 1980. pp. 654-661.
ed.
137. Tukey, J. W. Exploratory Data Analysis.
Addison-Wesley, Reading, Massachusetts,
1977. 688 pp.
138. U.S./Canada. Memorandum of Intent on
Transboundary Air Pollution. Impact Assess-
ment, Working Group I. Final Report, January
1983.
139. U.S. Environmental Protection Agency. Mbth-
ods for Chemical Analysis of Water and
Wastes. EPA-600/4-79-020, Cincinnati, Qhio,
1983.
140. Velleman, P. F., and D. C. Hoaglin. App ica-
tions. Basics and Computing of Exploratory
Data Analysis. Duxbury Press, Boston, Massa-
chusetts, 1981.
141. Wetzel, R. G. Limnology. 2nd Edition. Siiun-
ders College Publishing, Philadelphia, Penn-
sylvania, 1983. 767 pp.
142. Winter, T. C. Classification of Hydrologic Set-
tings of Lakes in the North Central United
States. Water Resources Research, 13(4):753-
767, 1977.
143. Wright, R. F. Predicting Acidification of North
American Lakes. Acid Rain Research Report
#4, NIVA, Oslo, 1983. 165 pp.
144. Zimmerman, A. P., and H. H. Harvey. Sensitiv-
ity to Acidification of Waters of Ontario and
Neighboring States. Final Report to Ontario
Hydro, University of Toronto, Toronto, On-
tario, Canada, 1979. 118 pp.
129
-------
Section 9
Glossary
B.I Abbreviations
A—estimated total lake area in a target
population
SA—area in a population of target lakes
Ac—estimated lake area below Xc
Acu—95% upper confidence limit on Ac
ANC—acid neutralizing capacity
cm—centimeter
DIG—dissolved inorganic carbon
DOC—dissolved organic carbon
DQO—data quality objective
ELS-I—Eastern Lake Survey - Phase I
EPA—Environmental Protection Agency
ERL-C—Environmental Research Laboratory -
Corvallis
EMSL-LV—Environmental Monitoring Systems
Laboratory - Las Vegas
F(x)—cumulative frequency distribution for
fakes
G(x)—cumulative areal distribution for lakes
gc—reference proportion (lake area)
ID—identification number
m—meter
M—median value of F(x)
MIBK—methyl isobutyl ketone
mg L~J—milligrams per liter
N—estimated number of lakes in a target
population
N*—number of lakes in frame population
n*—number of lakes selected in probabil-
ity sample
n**—number of lakes visited
n***—number of target lakes sampled
n'—effective sample size
n0—number of lakes not visited
nna—number of non-target lakes in the
sample determined after visiting the
lake
nnb—number of non-target lakes in the
sample determined from maps be-
fore visitation
Nc—estimated number of lakes below Xc
Ncu—95% upper confidence limit on Nc
NAPAP—National Acid Precipitation Assess-
ment Program
NBS—National Bureau of Standards
NLS—National Lake Survey
NSWS— National Surface Water Survey
NTU — nephelometric turbidity unit
ORNL — Oak Ridge National Laboratory
pc — reference proportion (number of
lakes)
PCU — platinum cobalt units
q — proportion of lakes visited
Q! — first quintile of F(x) or G(x)
Q2 — second quintile of F(x) or G(x)
Q3 — third quintile of F(x) or G(x)
Q4 — fourth quintile of F(x) or G(x)
QA — quality assurance
QC — quality control
QCCS — quality control check samples
RDL — required detection limit
%RSD — percent relative standard deviation
SAS — Statistical Analysis System
SD— standard deviation
SDL — system decision limit
SE — standard error
L~1 — microequivalents per liter
L~1 — micrograms per liter
— micrometers
(iS — microsiemens
USGS— United States Geological Survey
V — variance
W — weight or expansion factor
Xr — reference value of the variable X
9.2 Definitions
ACID NEUTRALIZING CAPACITY (ANC) - the total
acid-combining capacity of a water sample
determined by titration with a strong acid.
ANC includes ALKALINITY (carbonate spe-
cies) as well as other basic species (e.g., bo-
rates, dissociated ORGANIC ACIDS, alumino-
hydroxy complexes).
ACIDIC DEPOSITION - rain, snow, or dry fallout
containing high concentrations of sulfuric
and/or nitric acids, usually produced by at-
mospheric transformation of the by-products
of fossil fuel combustion (power plants,
smelters, autos, etc.); precipitation with a pH
<5.0 is generally considered to be unnatu-
rally acidic.
130
-------
ACIDIC LAKE - for this report, a lake with
NEUTRALIZING CAPACITY (ANC) s
ACIDIFICATION - loss of acid neutralizing
ity, either temporary or permanent.
capac-
that
used
has
lard
with
AIR-EQUILIBRATED - a sample aliquot
been brought to equilibrium with standard
air (300 ppm CO2) prior to analysis;
some pH and DIG measurements.
ALIQUOT - portion of a sample prepared
analysis of particular constituents,
separate container to the ANALYTICAI
ORATORY.
sent
or the
in a
LAB-
econ-
ALKALINITY - the titratable base of a samp
taining hydroxide, carbonate, and bicarbon-
ate ions, i.e. the equivalents of acid required
to neutralize the basic carbonate cpmpo-
nents.
a tea
de-
SUR-
reflect
ALKALINITY MAP CLASS - a geographic
fined by the expected ALKALINITY o
FACE WATERS (does not necessarily
measured alkalinity); used as a STRAtlFICA-
TION FACTOR in ELS-I design.
ALLOCTHONOUS - from a source outside the
system under consideration.
AMONG-BATCH PRECISION - the estimate bf pre-
cision that includes effects of different labo-
ratories and day-to-day difference wthin a
single laboratory.
, a lab-
ana-
the
to
ANALYTICAL LABORATORY - in this report
oratory under contract with the EPA
lyze ELS-I water samples shipped fro^m
FIELD LABORATORIES.
ANION - a negatively charged ion.
ANION-CATION BALANCE - in an electrical y neu-
tral solution, such as water, the total charge
of positive ions (CATIONS) equals the total
charge of negative ions (ANIONS); this calcu-
lation is used as a quality assurance check.
ANION DEFICIT - the amount in (xeq L~1 of mea-
sured ANIONS less than measured
CATIONS.
ANTHROPOGENIC - from a human sources
AUDIT - an onsite evaluation of a laboratory to
verify that standardized protocols arej being
followed.
AUDIT SAMPLE - a standardized water siample
submitted to laboratories to check overall
performance in sample analysis. Natural au-
dit samples were lake water; synthetic audit
samples were prepared by diluting concen-
ACID
trates of known composition. See FIELD
AUDIT SAMPLE and LABORATORY AUDIT
SAMPLE.
BASE CATION - a non-protolytic cation that does
not affect ANC; usually Ca+2 or Mg+2.
BATCH - a group of samples processed and ana-
lyzed together. A field batch of samples is
defined as all samples (including quality as-
surance samples) processed at one FIELD
LABORATORY in one day. A laboratory batch
is defined as all samples processed and ana-
lyzed at one ANALYTICAL LABORATORY, as-
sociated with one set of laboratory QUALITY
CONTROL samples.
BIAS - the expected value of the difference be-
tween a measurement and a true value being
measured.
BINOMIAL FORMULATION - refers to the binom-
ial probability distribution, which describes
the outcome of a sequence of identical and
independent random trials, each of which
can result in one of two outcomes (coin flip-
ping is a common example).
BLANK - a sample of ASTM Type I reagent grade
water analyzed as a QUALITY ASSURANCE/
QUALITY CONTROL sample in the ELS-I. See
FIELD BANK SAMPLE and LABORATORY
BLANK SAMPLE.
BUFFERING CAPACITY - the property of a solu-
tion that permits the relative concentrations
of hydrogen and hydroxyl ions to be main-
tained by neutralizing, within limits, added
acids or bases.
CARBONATE SYSTEM - a lake in which the major
part of the ANC is composed of carbonates;
organic or other weak anions are less than
10% of total ANION charge.
CATION - a positively charged ion.
CATION EXCHANGE - a process occurring in soil
in which acid cations (usually hydrogen ions)
are adsorbed and BASE CATIONS are re-
leased.
CHELATOR - a class of compounds, organic and
inorganic, that can bind metal ions and
change their biological availability.
CHROMOPHORE - a chemical group that gives
rise to color in a molecule.
CLEARWATER LAKE - for this report, a lake hav-
ing true color less than or equal to 30 PLAT-
INUM COBALT UNITS (PCU).
CLOSED LAKE - a lake with a surface water inlet
but no surface water outlet.
131
-------
CLOSED SYSTEM - method of measurement in
which a water sample is collected and ana-
lyzed for pH and DIG without exposure to the
atmosphere.
CLUSTER ANALYSIS - a multivariate classifica-
tion technique for identifying similar (or dis-
similar) groups of observations.
COLLOIDAL PARTICLES - particles of suspended
matter that are too small to settle from solu-
tion and often too small to be removed by
filtration.
CONDUCTANCE - a measure of the electrical con-
ductance (the reciprocal of the electrical re-
sistance) or total ionic strength of a water
sample.
CONFIDENCE BOUND or LIMIT (95%) - a value
that, in association with a statistic, has a 95%
chance of being above (or below) the true
value of the population parameter of interest.
CUMULATIVE FREQUENCY DISTRIBUTION - a
curve such that at any value x, the F(x) repre-
sents the estimated proportion of lakes in the
population having a value for that variable
that is less than or equal to x.
DARKWATER LAKE - for this report, a lake with
true color greater than 30 PLATINUM
COBALT UNITS (PCU).
DATA BASE - all computerized results of the Sur-
vey, which includes DATA SETS 1-4 as well
as back-up and historical data sets.
DATA FILE - a subset of information, often in tab-
ular form (e.g. the lake location file contains
information on lake identifiction and location
only).
DATA QUALIFIER - see FLAG and TAG.
DATA QUALITY OBJECTIVES (DQOs) - precision
limits desired by data users, established
prior to the beginning of the sampling effort.
DATA SET 1 - set of files containing RAW data.
DATA SET 2 - set of files containing VERIFIED
data.
DATA SET 3 - set of files containing VALIDATED
data.
DATA SET 4 - set of files containing final, en-
hanced data; missing values or errors in the
VALIDATED data set were replaced by substi-
tution values, duplicate values were aver-
aged, and negative measurements (except
ANC) were set equal to zero.
DEGASSING - the loss of dissolved C02 from a
water sample, which can result in changes in
pH and DIC between the time of collection
and the time of analysis.
DISSOLVED INORGANIC CARBON (DIC) - a mea-
sure of the dissolved carbon dioxide, car-
bonic acid, and carbonate and bicarbonate
anions which comprise the major part of
ANC in a lake.
DISSOLVED ORGANIC CARBON (DOC) - the or-
ganic fraction of carbon in a water sample
that is dissolved or unalterable (for this re-
port, 0.45 urn pore size).
DRAINAGE BASIN - see WATERSHED.
DRAINAGE LAKE - a lake with surface water out-
lets) or with both inlets and outlets.
EFFECTIVE SAMPLE SIZE (n') - within each STRA-
TUM, the total sample size modified because
of incomplete visitation.
EMPIRICAL DISTRIBUTION - observed distribu-
tion of results from the Survey.
EPILIMNION - the surface layer of warmer water
above the thermocline in a STRATIFIED
LAKE.
EQUIVALENT - unit of ionic charge; the quantity
of a substance that either gains or loses 1
mole of protons or electrons.
EXPANSION FACTOR - see WEIGHT.
EXTRACTABLE ALUMINUM - operationally de-
fined aluminum fraction that is extracted by
the procedure used in the ELS-I; this mea-
surement is intended to provide an indica-
tion of the concentration of aluminum that
may be available in a toxic form.
FALL CIRCULATION - see TURNOVER.
FIELD AUDIT SAMPLE - a standardized water
sample submitted to FIELD LABORATORIES
to check overall performance in sample anal-
ysis by both FIELD and ANALYTICAL LABO-
RATORIES. Natural field audit samples were
lake water; synthetic field audit samples
were prepared by diluting concentrates of
known composition.
FIELD BLANK SAMPLE - a sample of ASTM Type I
reagent grade water prepared at the lake site
by each field crew each day that sampling
occurred; these samples were analyzed by
both FIELD and ANALYTICAL LABORATO-
RIES.
FIELD DUPLICATE SAMPLE - a second sample of
lake water collected by the same crew at the
same lake site immediately after the ROU-
132
-------
TINE SAMPLE in accordance with standard-
ized protocols.
FIELD LABORATORY - mobile laboratory located
at each FIELD STATION in which sample
processing and measurement of selec ed
parameters were performed.
FIELD STATION - a location providing supportlfor
helicopters, sampling personnel, and FIEiLD
LABORATORIES during field sampling oper
ations.
FLAG - qualifier of a data value assigned during
the VERIFICATION and VALIDATION prc}ce-
dures.
FORM 1 - the field data form completed by field
sampling personnel.
FORM 2 - the field BATCH form completed by
field laboratory personnel.
FORM 3 - sample tracking form completed in the
field laboratory and used by the sample man-
agement office.
FRAME - a structural representation of a popula-
tion providing a sampling capability.
FRAME POPULATION - within each STRATUM
the ordered list of lakes depicted on
1:250,000-scale U.S. Geological Survey rrteips
GROUNDWATER - water in the part of the ground
that is completely saturated.
HARDWATER - water with high concentration^ of
calcium and magnesium salts.
HECTARE - 2.47 acres or ten thousand square
meters.
HIGH DOC LAKE - for this report, a lake with bOC
concentration greater than or equal to 6 mg
L-1.
HIGH SULFATE LAKE - for this report, a lake vi/ith
su If ate concentration greater than orequcl to
150 (jieq L"1.
ter
ED
the
HYPOLIMNION - the bottom layer of cold w;
below the THERMOCLINE in a STRATIF
LAKE.
IN SITU - referring to measurements made at
lake sampling site.
INDEX - in this report, one sample per lake, used
to represent chemical conditions in that Icike
INTERQUARTILE RANGE - the difference be-
tween the upper and lower quartile values in
a population frequency distribution (see
QUARTILE).
INSTRUMENT DETECTION LIMIT - for ei»ch
chemical parameter, a value calculated from
LABORATORY BLANK SAMPLES that indi-
cates the minimum concentration reliably
detectable by the instrument(s) used.
INTRA LABORATORY PRECISION GOAL - a preci-
sion goal based on the DATA QUALITY OB-
JECTIVES for the analysis of laboratory du-
plicates within a single laboratory.
IONIC STRENGTH - a measure of the interionic
effect resulting from the electrical attraction
and repulsion between various ions. In very
dilute solutions, ions behave independently
of each other and the ionic strength can be
calculated from the measured concentra-
tions of anions and cations present in the
solution.
ISOTHERMAL - defined as a temperature differ-
ence in a lake of less than 4°C between the
reading at 1.5 m below the surface and at 1.5
m above lake bottom (synonymous with un-
stratified or mixed).
LABORATORY AUDIT SAMPLE - a standardized
water sample submitted blind to ANALYTI-
CAL LABORATORIES to check overall perfor-
mance in sample analysis. All laboratory au-
dit samples were synthetic, prepared by
diluting concentrates of known chemical
composition.
LABORATORY BLANK SAMPLE - a sample of
ASTM Type I reagent grade water prepared and
analyzed by ANALYTICAL LABORATORIES
LABORATORY DUPLICATE SAMPLE - an aliquot
that is split and analyzed twice for a given
suite of variables at the ANALYTICAL LABO-
RATORIES.
LAKE ID - a unique, seven-character identification
code given to each lake in the Survey desig-
nating the REGION, SUBREGION, and ALKA-
LINITY MAP CLASS to which the lake be-
longs; e.g., lake ID IA2-101 designates the
101st lake in alkalinity map class 2 of Subre-
gion A in Region 1.
LAKE TYPE - a classification of lakes based on the
presence or absence of inlets, outlets, and
dams as represented on large-scale maps.
LARGE-SCALE MAPS - 1:24,000-, 1:25,000-, or
1:62,500-scale U.S. Geological Survey topo-
graphic maps.
LONG-TERM ACIDIFICATION - permanent loss of
ANC from a lake.
733
-------
LONG-TERM MONITORING PROGRAM - an EPA
sponsored program that monitors the chem-
istry of selected lakes and streams at least
three times a year.
LORAN-C GUIDANCE SYSTEM - a navigational
system based upon the time-delay of re-
ceived radio signals. The system provides
latitude and longitude at any point on the
earth's surface.
LOW ANC LAKE - for this report, a lake with ANC
less than or equal to 50 n,eq L~1.
LOW CALCIUM LAKE - for this report, a lake with
calcium concentration less than or equal to
50 (jueq L~1.
LOW DOC LAKE - for this report, a lake with DOC
concentration less than or equal to 2 mg L~1.
LOW pH LAKE - for this report, a lake with pH less
than or equal to 5.0.
MACROPHYTE - macroscopic forms of aquatic
vegetation, including macroalgae, aquatic
mosses and ferns, as well as true angio-
sperms.
MAP POPULATION - see FRAME POPULATION.
MATRIX - the physical and chemical makeup of a
sample being analyzed.
MEDIAN - (M) the value of x such that F(x) or
G(x) = 0.5; half the lakes in the population
are estimated to have x s MF and half the
lake area is estimated to have a value of
x
-------
QUALITY CONTROL CHECK SAMPLES (QCC5) -
samples of known concentration used to ver-
ify continued calibration of instruments
QUANTITATION LIMIT - for each chem cal
parameter, a value calculated from FIEiLD
BLANK SAMPLES that represents the lowest
concentration that can be measured with ea-
sonable precision.
QUARTILE - any of the three values that d ide
the population of a frequency distribution
into four equal classes, each representing 25
percent; used to measure range or variat on.
QUINTILE - any of the four values (Qv Q2, Q3, Q4)
that divide the population of a frequency dis-
tribution into five equal classes, each repre-
senting 20 percent; used to provide addi-
tional values to compare among populations
of lakes (see CUMULATIVE FREQUE CY
DISTRIBUTION).
RAW DATA SET - the initial data set (DATA S JT1)
that has received a cursory review to confirm
that data are provided in proper format nd
are complete and legible.
REFERENCE VALUE (Xc) - a concentration o in-
terest for a given chemical variable.
REGION - a major area of the conterminous
United States where a substantial number of
lakes with ALKALINITY less than 400 n-er- ~1
can be found.
REGULAR LAKE - in this study, a lake that is i the
PROBABILITY SAMPLE of lakes sele ed
from the FRAME POPULATION.
REMOTE BASE SITE - location serving as a ase
of operations for sampling crews wording
more than 100 miles from the FIELD LABO-
RATORY; samples were flown to the eld
laboratory daily.
REQUIRED DETECTION LIMIT - for each cherJi cal
parameter, the INSTRUMENT DETECTION
LIMIT required in the ANALYTICAL LABQRA-
TORY contract.
ROOT MEAN SQUARE - a summary statistic
culated as (ZSi/n2)1/2, where Si/n = the SJfi
ARD DEVIATION.
ROUTINE SAMPLE - the first lake sample
lected at a site in accordance with stand
ized protocols.
SAS - Statistical Analysis System, Inc. (Gary,
A statistical data file manipulation pac
that has data management, statistical,
graphical analysis abilities. The ELS-I
base was developed and analyzed prim
MD-
ol-
rd-
C).
ge
nd
ata
ily
using SAS software, and is distributed in
SAS format.
SECCHI DISK - a 20-cm-diameter, black and white
disk used to measure water transparency.
SECONDARY VARIABLES - chemical variables
measured during the Survey considered to
be important in providing additional data in
quantifying the chemical status of lakes, e.g.,
sodium, magnesium, potassium, nitrate,
chloride and total aluminum.
SEEPAGE LAKE - a lake with no permanent sur-
face water inlets or outlets.
SHALLOW LAKE - a lake from which a clean sam-
ple could not be obtained at 1.5 m below the
surface but could be obtained at 0.5 m; these
lakes were generally less than 3 m deep.
SMALL-SCALE MAP - 1:250,000-scale U.S. Geo-
logical Survey map.
SOFTWATER - water with low concentrations of
calcium and magnesium salts.
SPECIAL INTEREST LAKE - in this Survey, a lake
selected non-randomly, thus not part of the
PROBABILITY SAMPLE; selection was based
on quality and amount of data available or
involvement in other research programs.
SPLIT SAMPLE - a subsample of a field BATCH
sample that was sent to other laboratories
for analysis.
STANDARD DEVIATION - the square root of the
variance of a given statistic.
STRATIFICATION FACTORS - factors used to de-
fine STRATA prior to lake selection; the fac-
tors used in the ELS-I where REGION, SUB-
REGION and ALKALINITY MAP CLASS.
STRATIFIED DESIGN - a statistical design in
which the population is divided into strata,
and a sample selected from each STRATUM.
STRATIFIED LAKE - in this report, a lake with a
temperature difference greater than 4°C be-
tween the water layers at 1.5 m below the
surface and 1.5 m above the lake bottom. If
the temperature difference is also greater
than 4°C between the water layers at 1.5 m
below the surface and 60% of site depth, then
the lake is strongly stratified, if not, it is
weakly stratified.
STRATUM - in this Survey, a subpopulation of
lakes within an ALKALINITY MAP CLASS
within a SUBREGION and within a REGION,
as defined by the STRATIFIED DESIGN.
135
-------
SUBPOPULATION - any defined subset of the
TARGET POPULATION.
SUBREGIONS - areas within each REGION that
are similar in water quality, physiography,
vegetation, climate, and soil; used as a
STRATIFICATION FACTOR in ELS-I design.
SURFACE WATER - streams and lakes.
SYNOPTIC - relating to or displaying conditions
as they exist simultaneously over a broad
area.
SYSTEM DECISION LIMIT - for each chemical
parameter, a value calculated from FIELD
BLANK data that reliably (with 95% confi-
dence) indicates a concentration above back-
ground levels.
SYSTEMATIC ERROR - a consistent error intro-
duced in the measuring process. Such error
commonly results in biased estimations.
SYSTEMATIC RANDOM SAMPLING - a sampling
technique in which the units in the popula-
tion are ordered, a first sampling unit is ran-
domly drawn from the first K units, and every
kth unit afterward is included in the sample
(K being equal to N divided by the sample
size).
TAG - code on a data point that is added at the
time of collection or analysis to qualify the
data.
TARGET POPULATION - in this Survey, the lake
population of interest that was sampled. This
population was defined by the sampling pro-
tocol.
THERMOCLINE - the area of most rapid tempera-
ture change with depth in a STRATIFIED
LAKE.
TITRATION DATA - individual data points from
the modified Gran analysis of ANC.
TOPOGRAPHIC MAP - a map showing contours
of surface elevation.
TRANSPARENCY - the clarity of unfiltered water.
TRAILER DUPLICATE - a sample that is split and
analyzed twice in the FIELD LABORATORY.
TRUE COLOR - the color of water that has been
filtered or centrifuged to remove particles
that may impart an apparent color; true color
ranges from clear blue to blackish-brown.
little or no differences in temperature are ob-
served with depth.
VALIDATION - process by which data are evalu-
ated for quality with reference to the in-
tended data use; includes identification of
OUTLIERS and evaluation of potential SYS-
TEMATIC ERROR.
VERIFICATION - process of ascertaining the qual-
ity of the data in accordance with the mini-
mum standards established by the QUALITY
ASSURANCE plan.
WATERSHED - the geographic area from which
surface water drains into a particular lake.
WATERSHED DISTURBANCE - a disturbance of
the natural environment in a watershed
within 100 m of the shore as noted by field
samplers. Disturbances were roads, houses,
logging, mining and livestock.
WEAK ACID SYSTEM - a lake in which more than
10% of theanion charge results from organic
acids or other weak acid anions.
WEIGHT - the inverse of a sample lake's inclusion
probability; each sample lake represents this
number of lakes in the population.
WITHIN-BATCH PRECISION - the estimate of pre-
cision expected in the analysis of samples in
a BATCH by the same laboratory on any sin-
gle day (in this report, overall within-batch
precision includes the effects of sample col-
lection, processing and analysis; analytical
within-batch precision includes the effects of
sample analysis within ANALYTICAL LABO-
RATORIES).
TURBIDITY - a measure of light scattering by sus-
pended particles in an aqueous solution.
TURNOVER - a period of circulation in lakes when
136
U.S.GOVERNMENTPRINTINGOmCE:1986 •6'*6-ll&r 1*061 2REGIONNO. 4
-------
Upper Midwest
Northeast
Upper Peninsula of Michigan (2B)
Northcentral Wisconsin (2C)
Upper Great Lakes Area (2D)
:ral New England (1C)
Southern New England (1D)
Regions and Subregions, Eastern Lake Survey-Phase I
------- |