United States     Office of Acid Deposition, Environmental  EPA/600/3-86/054a
PA"600/3-86-054a C. 1 Environmental Protection Monitoring and Quality Assurance    January 1987
            Agency        Washington DC 20460
v-xEPA
            Research and Development
Western Lake Survey
Phase I

Characteristics of
Lakes in the Western
United States: Volume I.

Population
Descriptions and
Physico-Chemical
Relationships

-------
 o  S
 o  k-
 o  o
Q- -Z.
 o
O

-------
                                   EPA/600/3-86/054a
                                         January 1987
          Western  Lake Survey
                    Phase I

    Characteristics  of Lakes in the
         Western United States

  Volume I:  Population Descriptions
                       and
   Physico-Chemical Relationships
                      by

    D. H. Landers, J. M. Eilers, D. F. Brakke, W. S. Overton, P. E. Kellar,
 M. E. Silverstein, R. D. Schonbrod, R. E. Crowe, R. A. Linthurst, J. M. Omernik,
               S. A. league, and E. P. Meier
                 A Contribution to the
         National Acid Precipitation Assessment Program
           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 89193
                                  : T-onmental Protection Agency
                                   Library (5PL-16)
                                  ••-uborn Street, Room 1670
                                  ,L   60604

-------
                                  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 to
Lockheed  Engineering and Management Services Company, Inc., Nos.  68-
02-7288 and 68-02-3994 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:

Landers, D. H.1, J. M.  Eilers2,  D. F.  Brakke3,  W. S.  Overton4, P. E.  Kellar5,
  M. E.  Silverstein6, R. D. Schonbrod7, R. E. Crowe7, R. A. Linthurst8, J. M.
  Omernik9, S.  A. league2, and E. P. Meier7. 1987. Characteristics of Lakes
  in the Western United States. Volume I. Population Descriptions and Physico-
  Chemical  Relationships.  EPA/600/3-86/054a,  U.S.  Environmental
  Protection Agency, Washington, DC. 176 pp.

Eilers, J. M., P. Kanciruk10,  R.  A. McCord10, W. S. Overton,  L Hook11, D. J.
  Blick2, D. F. Brakke, P. E. Kellar, M. S. DeHaan2, M. E. Silverstein, and D. H.
  Landers. 1987. Characteristics of Lakes in the Western United States. Volume
  II. Data Compendium for Selected  Physical and Chemical Variables. EPA/
  600/3-86/054b, U.S. Environmental Protection Agency, Washington,  DC.
  425 pp.

Inquiries regarding the availability of the Western 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.
 'USEPA, Environmental Research Laboratory, 200 S W 35th Street, Corvallis, Oregon 97333
 2Northrop Services, Inc., Environmental Research Laboratory, 200 S W 35th Street, Corvallis, Oregon 97333
 3Western Washington University, Institute for Watershed Studies, Environmental Sciences Building, Room
  604, Belhngham, Washington 98225.
 "Oregon State University, Department of Statistics, Kidder Hall No 8, Corvallis, Oregon 97331.
 6Radian Corporation, 3200 E Chapel Hill Road, Research Triangle Park, North Carolina 27709
 6Lockheed Engineering and Management Services Company, Inc. 1050 E  Flamingo Road, Suite 120, Las
  Vegas, Nevada 89119
 'USEPA, Environmental Monitoring Systems Laboratory, 944 E Harmon Avenue, Las Vegas, Nevada 89193
 HJSEPA, Office of Research and Development, 401 M Street, S.W, Washington DC 20460. Present Address-
  Environmental Monitoring Systems Laboratory, Mail-Drop 39, Research Triangle Park, North Carolina 27711
 'USEPA, Environmental Research Laboratory, 200 S W 35th Street, Corvallis, Oregon 97333
 '"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-ACO5-84OR21400 for
  the U S. Department of Energy
 "Science Applications International Corporation, 800 Oak Ridge Turnpike, Oak Ridge, Tennessee 37831

-------
                               Preface
The National Surface Water Survey (NSWS) was begun in 1983 by the U.S.
Environmental Protection Agency (EPA) under the National Acid Precipitation
Assessment Program, a Federal Interagency Task Force mandated by Congress
in 1980. The goals of the NSWS are to: (1) determine the present chemical
conditions  in surface waters  in regions of the United  States considered
potentially sensitive to the effects of acidic deposition, (2) evaluate the temporal
variability  of chemistry and quantify the biological status of these waters,
and (3)  develop and implement a  project to monitor changes  over time  in
representative systems. The NSWS is one of several  major projects  in the
Acid Deposition Aquatic Effects Research Program. This program, one of many
research programs investigating acidic deposition, is administered in the Acid
Deposition  and Atmospheric Research Division: Office of Acid Deposition,
Environmental Monitoring and Quality Assurance in the U.S. EPA Office  of
Research and Development.

The Aquatic Effects Research  Program addresses four primary  assessment
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?


Five major research projects within  the Aquatic Effects Research  Program
specifically address  these assessment  questions from  a regionalized
perspective. These projects and their goals are:

 1. National Surface  Water  Survey  (NSWS):  to determine  the present
    chemistry, characterize the chemical temporal variability, and determine
    the key biological resources of lakes and streams in potentially sensitive
    regions of the United States;

 2. Direct/Delayed Response Project (DDRP): to predict changes  in these
    aquatic resources at various levels of acidic deposition, considering the
    terrestrial and aquatic variables that influence these changes;

 3. Watershed Manipulation Project (WMP): to verify that predictions of future
    change are reasonably sound by manipulating watershed  catchments
    or system components;

 4. Episodic Response Project (ERP): to evaluate the  regional importance
    of short-term acidification resulting from episodic hydrologic events and
    its  effect on the quality of the biological environment; and

-------
5.  Long-Term  Monitoring  Project (LTMP):  to test the validity of predicted
    changes through long-term monitoring of regionally characteristic lakes
    and streams.

The NSWS, including surveys of lakes and streams, addresses the first goal
of the Aquatic  Effects Research Program. Understanding the  national-scale
effects of acidic deposition on aquatic resources  requires that  the present
chemical status of surface waters be understood on large geographical scales.
The Western  Lake  Survey - Phase I  (WLS-I) was  designed  to describe
statistically the present surface water chemistry on a  regional scale  in the
western United States. Although the cause-and-effect  relationship between
acidic deposition  and surface  water  response  cannot be determined on the
basis of the WLS-I data alone, continued analysis of these data will contribute
significantly to  furthering our  understanding of the extent to which western
lakes are at risk due to acidic deposition. Determining the relationships between
acidic deposition and surface water chemistry and biology is the goal of future
activities within the Aquatic Effects Research Program.

-------
                               Volume I
                               Contents

Section                                                            Page

Notice	ii
Preface	iii
Volume II Contents	x
Figures 	xi
Tables	xxi
Related Documents 	xxvi
Acknowledgments	xxvii
Executive Summary	xxxiii

  1.  Introduction	1

  2.  Methods 	3

     2.1  Design	3
          2.1.1   Data Quality Objectives	3
          2.1.2   Summary of the Statistical Design	3
          2.1.3   Definition of Study Area	4

     2.2  Lake Selection	4
          2.2.1   Probability Sample	4
          2.2.2   Identification of Non-target Lakes	5
          2.2.3   Identification of Wilderness Lakes  	8
          2.2.4   Special Interest Lakes	8
          2.2.5   Final Lake Lists and Maps  	12

     2.3  Design Applications and Restrictions	12
          2.3.1   Extrapolation from Sample to Population	12
          2.3.2   Estimating the Target Population Size and Attributes ... 12
          2.3.3   Restrictions	15

     2.4  Lake Characterization	15
          2.4.1   Lake Area	15
          2.4.2   Elevation  	15
          2.4.3   Hydrologic Lake Type	15
          2.4.4   Watershed Area	16
          2.4.5   Land Use/Land Cover	16
          2.4.6   Slope	16

     2.5  Sampling and Analytical Methods  	16
          2.5.1   Field Sampling Activities  	17
                 2.5.1.1  Site Description and Location	17
                 2.5.1.2  In Situ Measurements and Sample
                         Collection	17
          2.5.2   Field Laboratory Operations	19
          2.5.3   Analytical  Laboratory Operations	21

     2.6  Calibration Study	21

-------
                              Volume I
                       Contents (continued)

Section                                                           Page

    2.7   Data Base Management	22

 3.  Quality Assurance	26

    3.1   Design of the Quality Assurance Program 	26
         3.1.1  Selection of Analytical Laboratories	26
         3.1.2  Training of Sampling and Field Laboratory Crews	26
         3.1.3  Communications	26
         3.1.4  On-Site Inspections	26
         3.1.5  Quality Assurance/Quality Control Data
               Collection and Analysis  	26
                3.1.5.1  Quality Assurance Samples	27
                3.1.5.2  Field Sampling and Field Laboratory
                        Quality Control Samples	29
                3.1.5.3  Analytical Laboratory Quality Control
                        Samples	29

    3.2   Data Verification	29
         3.2.1  Review of Field Data Forms 	29
         3.2.2  Initial Review of Analytical Laboratory Sample
               Data Package	29
         3.2.3  Review of Quality Assurance/Quality
               Control Data	29

    3.3   Data Validation	30

    3.4   Development of Final Data Set	31

 4.  Quality Assurance and Calibration Study Results	32

    4.1   Site Confirmation	32

    4.2   Detection Limits (Detectability)	32

    4.3   Precision	33

    4.4   Effects of Holding Time on Sample Concentration	37

    4.5   Assessment of the Effect of Different Field Sampling
         Protocols	40
         4.5.1  Calibration of the Ground Sample Data  	40
         4.5.2  Relative Bias Between Analytical Laboratories	44

    4.6   Assessment of Data Quality	47

 5.  Results of Population Estimates	52

    5.1   Data Presentation and Considerations	52
         5.1.1  Presentation	52
         5.1.2  Design Considerations	54
                5.1.2.1  Using Weights	54
                5.1.2.2  Evaluation  of Alkalinity Map Classes	55

    5.2   Description of Target Population	55
         5.2.1  Number of Lakes Sampled	55

-------
                              Volume I
                       Contents (continued)

Section                                                           Page

         5.2.2  Treatment of Shallow, Thermally Stratified, and
               Large Lakes	56
         5.2.3  Description of Target Population and Sample	56

    5.3   Physical Characteristics of Lake Populations  	56

    5.4   Population Estimates for Primary Variables 	59
         5.4.1  Acid Neutralizing Capacity	59
                5.4.1.1   Reference Values 	59
                5.4.1.2   Subregional Estimates and Distribution of
                         Lakes Having Low Acid Neutralizing
                         Capacity	59
                5.4.1.3   State Estimates	60
         5.4.2  pH 	60
                5.4.2.1   Subregional Estimates	60
                5.4.2.2   State Estimates	60
         5.4.3  Sulfate	60
         5.4.4  Calcium	66
         5.4.5  Extractable Aluminum, Clearwater Lakes	67
         5.4.6  Dissolved Organic Carbon 	67

    5.5   Statistics for Population Descriptions, Primary Variables	67

    5.6   Statistics for Population Descriptions, Secondary Variables  .... 68
         5.6.1  Nutrients	68
         5.6.2  True Color, Turbidity, and Secchi Disk Transparency. . . .71
         5.6.3  Sodium, Potassium, and Magnesium	72
         5.6.4  Iron, Manganese, and Total Aluminum	72
         5.6.5  Other Secondary Variables	73

    5.7   Statistics for Lakes in Wilderness Areas and National Parks ... .73
         5.7.1  Wilderness Area and National Park Lakes	73
         5.7.2  Subpopulations in Wilderness Areas and National
               Parks	74

    5.8   Sample Statistics for Special Interest Lakes	76

    5.9   Statistics for Dilute Lakes	76

 6.  Associations Among Variables	85

    6.1   pH and Acid Neutralizing Capacity	85
         6.1.1  Comparison of pH Measurements	85
         6.1.2  Relationship  between pH and Acid Neutralizing
               Capacity	85

    6.2   Major  Cations and Anions 	86
         6.2.1  Individual Example Lakes	87
         6.2.2  Calcium and  Magnesium	88
         6.2.3  Calcium and  Sulfate	88
         6.2.4  Acid Neutralizing Capacity versus Base Cations	90
         6.2.5  Relative Abundance of Major Ions	90
         6.2.6  Relationships Among  Major Ions	95
         6.2.7  Dissolved Organic Carbon 	100
                6.2.7.1   Dissolved Organic Carbon and True Color  ... 100
                6.2.7.2   Anion Deficit	100
                                  vii

-------
                              Volume I
                       Contents (continued)

Section                                                         Page

    6.3  Geology	103

    6.4  Physical Lake Characteristics and Water Chemistry	105
         6.4.1   Lake Area	105
         6.4.2   Lake Elevation	107
         6.4.3   Hydrology	109
                6.4.3.1  Hydrologic Lake Type	109
                6.4.3.2  Hydraulic Residence Time	110

 7.  Subregional Characteristics	118

    7.1   California (4A)	118
         7.1.1   Sierra  Nevada	118
         7.1.2   Other Geomorphic Units	121

    7.2  The Pacific Northwest (4B)	122
         7.2.1   The North Cascades	124
         7.2.2   The Oregon Cascades	124
         7.2.3   Other Geomorphic Units	126
         7.2.4   Sulfate in Geomorphic Units in the Pacific
                Northwest 	126

    7.3  Northern Rocky Mountains (4C)	127
         7.3.1   Idaho Batholith	127
         7.3.2   Bitterroot Range	129
         7.3.3   Other Geomorphic Units	129

    7.4  Central Rocky  Mountains (4D)	129
         7.4.1   The Wind River Uplift	131
         7.4.2   The Uinta Arch	131
         7.4.3   Other Geomorphic Units	131

    7.5  Southern Rocky Mountains (4E)	134
         7.5.1   The Front Range	135
         7.5.2   Other Geomorphic Units	135

    7.6  Ionic Relationships for Lakes in the Major Geomorphic
         Units	138

 8.  Comparison of Results from the Western and Eastern Lake
    Surveys	140

    8.1   Physical Characteristics	140

    8.2  Primary Variables	140

    8.3  Associations Among Variables	145

    8.4  Discrete Subpopulations of Lakes in the East and West	146

 9.  Conclusions	149

10.  References 	150

11.  Glossary	157
                                  VIII

-------
                           Volume I
                     Contents (continued)

 Section                                                    Page

   11.1  Abbreviations	157
   11.2  Definitions	159

Appendix	169
                                IX

-------
                              Volume II
                              Contents

Section                                                          Page

Notice	ii
Preface	iii
Related Documents	vi
Figures	vii
Tables	xxv

1.  Introduction	1

2.  Population Estimates for Selected Physical and Chemical Variables . . .5

    2.1  Applications of the Design	5
          2.1.1  Extrapolation from Sample to Population	5
          2.1.2  Estimating the Target Population Size and
                Subpopulation	5

    2.2  Data Presentation	8

3.  USGS Maps Used to Define and Characterize the Lake
    Population	190

    3.1  Small-scale Maps	190
    3.2  Large-scale Maps	190

4.  Maps Showing Lake Locations by State	210

5.  Lakes Sampled and Individual Lake Data	221

    5.1  Probability Sample Lakes	221
          5.1.1  List of Probability  Sample Lakes Sorted by State and
                Lake Name	221
          5.1.2  List of Probability  Sample Lakes Sorted by Lake ID .... 221
          5.1.3  Data for Individual Probability Sample Lakes  	222

    5.2  Special Interest Lakes	223
          5.2.1  List of Special Interest Lakes Sorted by State and
                Lake Name	223
          5.2.2  List of Special Interest Lakes Sorted by ID 	223
          5.2.3  Data for Individual Special Interest Lakes	223

6.  References	424

-------
                               Figures


Number                                                           Page

 2-1  Subregions and alkalinity map classes ( •: < 100 //eq L~1;
      m\ 100-199 fjeq L"1; D: 200-400 /veq L~1), in study areas
      of the Western Lake Survey - Phase I 	6

 2-2  Locations of Forest Service wilderness areas ( B ) and national
      parks ( & ) containing lakes sampled in the California
      (4A) and Pacific Northwest (4B) subregions, in study areas
      of the Western Lake Survey - Phase I 	9

 2-3  Locations of Forest Service wilderness areas ( El) and national
      parks ( ^ ) containing lakes sampled in the Rocky Mountain
      subregions: the Northern Rockies (40), the Central Rockies
      (4D), and the Southern Rockies (4E), in study areas of the
      Western Lake  Survey - Phase I. Lakes located in roadless areas
      are also shown ( E3)	10

 2-4  Procedures used to estimate the target population size, Western
      Lake Survey -  Phase I  	13

 2-5  Field sampling activities. Western Lake Survey - Phase I. Field
      crews accessed most sampling sites in wilderness areas by
      ground and in  non-wilderness areas by helicopter 	18

 2-6  Field laboratory activities. Western Lake Survey - Phase I.
      Samples collected by field crews were processed at mobile
      field laboratories before shipment to analytical laboratories
      for analysis	20

 2-7  Allocation of samples to analytical laboratories for the calibration
      study. Western Lake Survey - Phase I. This procedure was used
      to assign routine (R), duplicate (D), and triplicate (T) samples
      collected by ground crews (GC) and helicopter crews (HC) to one
      of two analytical laboratories	23

 2-8  Procedures used to develop Data Set 1 (raw). Data Set 2  (verified).
      Data Set 3 (validated), and Data Set 4 (final) for the Western
      Lake Survey -  Phase I data base	24

 3-1  Origin, use, and flow of quality assurance/quality control
      samples, Western Lake Survey - Phase I. Results of the analyses
      of these samples were used to quantify data quality and to ensure
      that error was  minimized at each level of sample handling and
      processing 	28

-------
                       Figures (continued)


Number                                                         Page

 4-1  Results of ANC (/ueq L~1) measurements and pH determinations
      on air-equilibrated duplicate samples collected by helicopter
      and ground crews from 45 lakes for the calibration study
      during the Western Lake Survey - Phase I. The dimensions of the
      boxes indicate the magnitude of sampling and measurement
      error; sampling bias is evident when the position of the boxes
      consistently deviates from the 1:1 (diagonal) line	41

 4-2  Results of sulfate (/ueq L"1) and calcium (/ueq L"1) analyses on
      samples collected by helicopter  and ground crews from 45 lakes
      for the calibration study during the Western Lake Survey - Phase I.
      The dimensions of the boxes indicate the magnitude of sampling
      and measurement error; sampling bias is evident when the
      position of the boxes consistently deviates from the 1:1
      (diagonal) line 	42

 4-3  Results of extractable aluminum (/L/g L"1) and DOC (mg L~1)
      measurements on duplicate samples collected by helicopter and
      ground crews from 45 lakes for the calibration study during the
      Western Lake Survey - Phase I.  The dimensions of the boxes
      indicate the magnitude of sampling and measurement error;
      sampling bias is evident when the position of the boxes
      consistently deviates from the 1:1 (diagonal) line	43

 4-4  Results of ANC (/ueq L"1) and pH determinations on air-
      equilibrated samples performed by Lab 1 and Lab 2. Analyses
      were conducted on duplicate samples collected by helicopter
      and ground crews from 45 lakes for the calibration study during
      the Western Lake Survey - Phase I. The dimensions of the boxes
      indicate the magnitude of sampling and measurement error;
      laboratory bias is evident when  the position of the  boxes
      consistently deviates from the 1:1 (diagonal) line	46

 4-5  Results of sulfate (/ueq L""1) and calcium (/ueq L"1) measurements
      by Lab 1  and Lab 2. Analyses were conducted on duplicate
      samples collected by helicopter  and ground crews from 45 lakes
      for the calibration study during the Western Lake Survey -
      Phase I. The dimensions of the boxes indicate the magnitude
      of sampling and measurement error; laboratory bias is evident
      when the position of the boxes consistently deviates from the
      1:1 (diagonal) line	48

 4-6  Results of extractable aluminum (/ug L"1) and DOC (mg L"1)
      measurements performed by Lab 1 and Lab 2. Analyses were
      conducted on duplicate samples collected by helicopter and
      ground crews from 45 lakes for the calibration study during
      the Western Lake Survey - Phase I. The dimensions of the boxes
      indicate the magnitude of sampling and measurement error;
      laboratory bias is evident when  the position of the  boxes
      consistently deviates from the 1:1 (diagonal) line	49

 5-1  F(x) and G(x) distributions (	) and the 95 percent  upper
      confidence limits (	) on these distributions for ANC (//eq L"1)
      for the target population of lakes ( > 1 ha and < 2000 ha) sampled
      in Region 4 (West), Western Lake Survey -  Phase  I. The population

-------
                        Figures (continued)


Number                                                          Page

     statistics, quintiles (Qi-CU) and medians, and three reference values
     (Xci, XC2, and Xca) for ANC are also shown. The proportions
     (Pc or gc), and number or area (Nc or Ac) of lakes below the reference
     values, and their 95% upper confidence bounds (Ncu or Acu) are
     given below each distribution	53


5-2  Locations of lakes sampled in California (4A) with ANC (yueq L"1)
     in one of three classes ( • : ANC < 50, + : ANC > 50-200, and
     0 : ANC > 200), Western Lake Survey - Phase I. Geographic area
     covered by the target population is bounded by a dashed line	61

5-3  Locations of lakes sampled in the Pacific Northwest (4B) with ANC
     (/ueq L"1)  in one of three classes ( • : ANC < 50, + : ANC > 50-200,
     and 0 : ANC > 200), Western Lake Survey - Phase I. Geographic
     area covered by the target population is bounded by a dashed
     line	62

5-4  Locations of lakes sampled in the Northern Rockies (4C) with ANC
     (fjeq L"1)  in one of three classes ( • : ANC < 50, + : ANC > 50-200,
     and 0 : ANC > 200), Western Lake Survey - Phase I. Geographic
     area covered by the target population is bounded by a dashed
     line	63

5-5  Locations of lakes sampled in the Central Rockies (4D) with ANC
     (/ueq L"1)  in one of three classes ( • : ANC < 50, + : ANC > 50-200,
     and 0 : ANC > 200), Western Lake Survey - Phase I. Geographic
     area covered by the target population is bounded by a dashed
     line	64

5-6  Locations of lakes sampled in the Southern Rockies (4E) with ANC
     (/ueq L"1)  in one of three classes ( • : ANC < 50, + : ANC > 50-200,
     and 0 : ANC > 200), Western Lake Survey - Phase I. Geographic
     area covered by the target population is bounded by a dashed
     line	65

5-7  Cumulative frequency distributions [F(x>] for ANC (fjeq  L"1) and
     pH in California (	), the  Pacific Northwest (	), the
     Northern Rockies (	), the Central Rockies (	),  and the
     Southern Rockies (	), Western Lake Survey - Phase 1	69

5-8  Cumulative frequency distributions [F(x>] for calcium (/ueq L"1)
     and inverse cumulative frequency distributions [1 -F(x)] for
     sulfate (/ueq L"1) in California  (	), the Pacific Northwest
     (	), the Northern Rockies (	), the Central Rockies
     (	), and the Southern Rockies (	), Western Lake
     Survey -  Phase I  	70

5-9  Cumulative frequency distributions [F(x)J for ANC (/ueq  L"1) in
     wilderness area (	) and  non-wilderness area (	) lakes
     in California (4A), the Pacific  Northwest (4B), the Northern
     Rockies (4C), the Central Rockies (4D), and the Southern Rockies
     (4E), Western Lake Survey - Phase I 	75
                                 XIII

-------
                        Figures (continued)


Number                                                         Page


5-10  Cumulative frequency distributions [F(x)J for calcium (yueq L"1) in
      wilderness area (	) and non-wilderness area (	) lakes
      in California (4A), the Pacific Northwest (4B), the Northern
      Rockies (4C), the Central Rockies (4D), and the Southern Rockies
      (4E), Western Lake Survey - Phase I 	75

5-11  Inverse cumulative frequency distributions [1 -F(x)] for sulfate
      (yueq L"1) in wilderness area (	) and non-wilderness area
      (	) lakes in California (4A), the  Pacific Northwest (4B),
      the Northern Rockies (4C), the Central Rockies (4D), and the
      Southern Rockies (4E), Western Lake Survey - Phase I  	76

5-12  Concentrations (fjeq L"1) of ANC, calcium, and sulfate in lakes
      in selected wilderness areas and national parks. Western Lake
      Survey - Phase I. Numbers in parentheses after each area name
      are those shown in Figures 2-2 and 2-3 and Table 2-4 in Section
      2.2.3. First quintiles (Qi, 20th percentile), medians (M, 50th
      percentile), and fourth quintiles (Ch, 80th percentile), are
      given for each  variable (O	1	a )	77
                            Qi  M  Q4

5-13  Locations of dilute  lakes (conductance < 10 pS crrT1) sampled
      in California (4A), Western Lake Survey - Phase I. Geographic
      area covered by the target population is bounded by a dashed
      line	79

5-14  Locations of dilute lakes (conductance < 10 //S cm"1) sampled
      in the Pacific Northwest (4B), Western Lake Survey - Phase I.
      Geographic area covered  by the target population is bounded by
      a dashed line	80

5-15  Locations of dilute lakes (conductance < 10 jt/S cm"1) sampled
      in the Northern Rockies (4C), Western Lake Survey - Phase I.
      Geographic area covered  by the target population is bounded by
      a dashed line	81

5-16  Locations of dilute lakes (conductance < 10 pS cm"1) sampled
      in the Central  Rockies (4D),  Western Lake Survey - Phase I.
      Geographic area covered by the target population is bounded by
      a dashed line	82

5-17  Locations of dilute lakes (conductance < 10 /uS cm"1) sampled
      in the Southern Rockies (4E), Western Lake Survey - Phase I.
      Geographic area covered by the target population is bounded
      by a dashed line 	83

 6-1  Relationship between closed system pH and measured ANC for
      lakes with pH < 9.0 and ANC < 1000 /ueq L"1 in all subregions,
      Western Lake  Survey - Phase I. This same relationship is shown
      in the inset for air-equilibrated pH	86

 6-2  Ionic composition of one lake selected from each of the five
      western subregions surveyed during the Western Lake Survey  -
      Phase I. Each lake was selected on the basis of having concen-

                                  XIV

-------
                        Figures (continued)


Number                                                         Page

      trations of ANC (//eq L~1) and DOC (mg L~1) approximately equal to
      the first quintile (Qi, 20th percentile) values estimated for the
      subregional target population of lakes. Cations and anions are
      expressed as percentages of the total  measured ionic equivalents
      (//eq L"1);  ions with concentrations < 5 //eq  L~1 are not shown.
      Champion Lake (4C2-047) is shown on the  USGS topographic
      map as an unnamed lake and is listed as "No Name" in the data
      base and  lake lists, Volume II	87

 6-3  Relationship between calcium ( < 500 //eq  L~1) and magnesium
      ( < 250 //eq L~1), Western Lake Survey - Phase I  	88

 6-4  Median sulfate concentrations (//eq L~1) for  lakes in three
      measured ANC classes for California (4A), the Pacific Northwest
      (4B), the Northern Rockies (4C), the Central Rockies (4D), and
      the Southern  Rockies (4E), Western Lake Survey - Phase I	90

 6-5  Relationship between measured ANC ( < 200 //eq L~1) and the
      sum of calcium and magnesium (//eq L"1) for lakes in California
      (4A), Western Lake Survey - Phase I. The alkalinity map class
      from which each lake was selected is shown  ( Q < 100 //eq L"1;
      A: 100-199 //eq L'1; X: 200-400//eq L"1; Omernik and
      Griffith 1986)	91

 6-6  Relationship between measured ANC ( < 200 //eq L"1) and the
      sum of calcium and magnesium (//eq L"1) for lakes in the Pacific
      Northwest (4B), Western Lake Survey - Phase I. The, alkalinity
      map class from which each lake was selected is shown
      (D: < 100//eq L"1; A: 100-199//eq L'1; X: 200-400//eq L"1;
      Omernik and Griffith 1986)	91

 6-7  Relationship between measured ANC ( < 200 //eq L"1) and the
      sum of calcium and magnesium (//eq L~1) for lakes in the
      Northern  Rockies (4C), Western Lake Survey - Phase I. The
      alkalinity map class from which each lake was selected is shown
      (D:< 100//eq L"1; A: 100-199//eq L'1; X: 200-400//eq L"1;
      Omernik and Griffith 1986)	92

 6-8  Relationship between measured ANC ( < 200 //eq L"1) and the
      sum of calcium and magnesium (//eq L~1) for lakes in the
      Central Rockies (4D), Western Lake Survey - Phase I. The
      alkalinity  map class from which each lake was selected is shown
      ( D: < 100 //eq L"1; A: 100-1 99 //eq L"1; X: 200-400 //eq L"1;
      Omernik and Griffith 1 986)	92

 6-9  Relationship between measured ANC ( < 200 //eq L~1) and the
      sum of calcium and magnesium (//eq L~1) for lakes in the
      Southern Rockies (4E), Western Lake Survey  - Phase I. The
      alkalinity  map class from which each lake was selected is shown
      ( A: < 100 //eq L"1; A: 100-1 99 //eq L~1;  X: 200-400 //eq L"1;
      Omernik and Griffith 1 986)	93

6-10  Concentration of chloride (//eq L~1) versus distance ( < 100 km)
      from the Pacific Coast for lakes in the Klamath Mountains in
      California (4A) and from Puget Sound or the Pacific Ocean

                                 xv

-------
                        Figures (continued)


Number                                                          Page

       (whichever is less) in five geomorphic units in the state of
       Washington in the Pacific Northwest (4B), Western Lake
       Survey - Phase I. Values in the Pacific Northwest are shown for
       the Wenatchee Mountains ( • ), the Middle Washington
       Cascades ( A ), the North Washington Cascades ( + ),
       the Puget Lowlands ( X ), and the Olympic Mountains ( D ).
       Distance from the coast was determined from a SAS file (SAS
       Institute 1985)	94

 6-11  Concentration of sodium (/ueq L~1) versus distance (< 100 km)
       from Puget Sound or the Pacific Ocean (whichever is less) for
       lakes in five geomorphic units in the state of Washington in
       the Pacific Northwest (4B), Western Lake Survey - Phase I.
       Values are shown for the Wenatchee Mountains (• ), the Middle
       Washington Cascades ( A ), the North Washington Cascades
       ( + ),  the Puget Lowlands ( X ), and the Olympic
       Mountains ( D ). Distance from the coast was determined
       from a SAS file (SAS Institute 1985)	96

 6-12  Relationship between sodium ( < 100 fjeq L"1) and chloride
       ( < 50 fjeq L"1) for lakes in the Pacific Northwest (4B) and the
       Northern Rockies (4C), Western Lake Survey - Phase I. The
       solid line indicates the ratio of sodium to chloride in sea water	97

 6-13  Trilinear diagram showing the relative abundance of major
       anions and cations for lakes in California (4A), Western Lake
       Survey - Phase I. Ratios are expressed as percent of total
       ionic concentration, increasing from 0 to 100 percent along the
       axes in the direction of the arrows. Lakes for which the total
       ionic composition cannot be described as being dominated by
       calcium bicarbonate are circled and their lake ID numbers are
       given 	98

 6-14  Trilinear diagram showing the relative abundance of major
       anions and cations for lakes in the Pacific Northwest (4B),
       Western Lake Survey - Phase I. Ratios are expressed as percent
       of total ionic concentration, increasing from 0 to 100 percent
       along the axes in the direction of the arrows. Lakes for which
       the total ionic composition cannot be described as being dominated
       by calcium bicarbonate are circled and their lake ID numbers
       are given	98

 6-15  Trilinear diagram showing the relative abundance of major
       anions and cations for lakes in the Northern Rockies (4C),
       Western Lake Survey - Phase I. Ratios are expressed as percent
       of total ionic concentration, increasing from 0 to 100 percent
       along the axes in the direction of the arrows. Lakes for which
       the total ionic composition cannot be described as being dominated
       by calcium bicarbonate are circled and their lake ID numbers
       are given	99

 6-16  Trilinear diagram showing the relative abundance of major
       anions and cations for lakes the Central Rockies (4D), Western
       Lake Survey - Phase I. Ratios are expressed as percent of total

-------
                        Figures (continued)


Number                                                          Page

       ionic concentration, increasing from 0 to 100 percent along the
       axes in the direction of the arrows. Lakes for which the total
       ionic composition cannot be described as being dominated by
       calcium bicarbonate are circled and their lake ID numbers are
       given 	99

 6-17  Trilinear diagram showing the relative abundance of major
       anions and cations for lakes in the Southern Rockies (4E),
       Western Lake Survey - Phase I. Ratios are expressed as percent
       of total ionic concentration, increasing from 0 to 100 percent
       along the axes in the direction of the arrows. Lakes for which
       the total ionic composition cannot be described as being dominated
       by  calcium bicarbonate are circled and their lake ID numbers
       are given	100

 6-18  Relationship between DOC ( < 10 mg L"1) and true color ( < 50
       PCU) for lakes in all subregions. Western Lake Survey - Phase I... 101

 6-19  Relationship between the sum of anions ( < 200 /ueq L"1) and the
       sum of cations ( < 200 /ueq L"1) for lakes in California ( • ), the
       Pacific Northwest ( A ), the Northern Rockies ( + ), the Central
       Rockies ( X ), and the Southern Rockies  ( D ), Western Lake
       Survey - Phase I. Points falling below the 1:1  (diagonal) line
       indicate an anion deficit, and above the  line, a cation deficit.
       Inset shows the same relationship for anions < 2000 /ueq L"1
       and cations < 2000 /ueq L'1	102

 6-20  Relationship between DOC (mg L"1) and anion deficit (/ueq L"1) for
       lakes in all subregions with total  measured ionic equivalents
       < 500 /ueq L"1, Western Lake Survey - Phase I. Anion deficit is
       the difference (/ueq L"1) between the sum of measured anions
       and the sum of measured cations	104

 6-21  Percentage of low ANC ( < 50 /ueq L"1) lakes by bedrock class
       (Norton et al.  1982) in California  (4A), the Pacific Northwest (4B),
       the Northern Rockies (4C), the Central Rockies (4D), and the
       Southern Rockies (4E), Western Lake Survey - Phase I. Bedrock
       sensitivity is highest in Class 1 and lowest in Class 3	106

 6-22  Locations and names of geomorphic units used in data analysis for
       specific subpopulations of lakes,  Western Lake Survey - Phase  I.
       These units represent physiographic areas of common geologic
       origin and generally coincide with major mountain ranges	107

 6-23  Relationship between ANC ( < 1000 /ueq L"1) and lake area
       ( < 100 ha) for lakes in the Pacific Northwest (4B), Western Lake
       Survey - Phase I. The alkalinity map class from which each lake
       was selected  is shown ( D: < 100 /ueq L"1; A:  100-1 99 /ueq L"1;
       X:  200-400 /ueq  L"1; Omernik and Griffith 1 986)	108

 6-24  Estimated number of lakes with ANC <  50 /ueq L"1 or ANC > 50
       /ueq L"1 by lake area class, Western Lake Survey -  Phase I	109

 6-25  Concentrations (/ueq L~1) of ANC,  sulfate, and calcium in lakes
       < 4 ha (	) and > 4 ha (	) in California (4A), the Pacific
                                 XVII

-------
                        Figures (continued)


Number                                                          Page

      Northwest (4B), the Northern Rockies (4C), the Central Rockies
      (4D), the Southern Rockies (4E), and the West (4) in the Western
      Lake Survey - Phase I. First quintiles (Qi, 20th percentile),
      medians (M, 50th percentile), and fourth quintiles (Q4, 80th
      percentile) are given for each variable (O	1	D)	110
                                         Qi  M  Q4

6-26  Cumulative frequency distributions [F(x>] for ANC ( < 1000 //eq L"1)
      and calcium (< 1000//eq L~1), and inverse cumulative frequency
      distributions [1-F(x>] for sulfate ( < 100/ueq L~1) in lakes < 4 ha
      (	) and > 4 ha (	) in the Northern Rockies (4C),
      Western Lake Survey - Phase I	111

6-27  Relationship between ANC ( < 1000 //eq L"1) and elevation (m)
      for lakes in California (4A), Western Lake Survey - Phase I.
      The alkalinity map class from which each lake was selected is
      shown ( D: < 100 yueq L"1; A: 100-199 //eq L"1; X: 200-400
      //eq L"1; Omernik and Griffith 1986)	112

6-28  Relationship between ANC ( <  1000 //eq L~1) and elevation (m)
      for lakes in the Pacific Northwest (4B), Western  Lake Survey -
      Phase I. The alkalinity map class from which each lake was
      selected is shown ( D: < 100 //eq  L~1; A: 100-199 //eq L"1;
      X: 200-400 //eq L"1; Omernik and Griffith 1986)	113

6-29  Relationship between ANC ( <  1000 //eq L"1 ) and elevation (m)
      for the Northern Rockies (4C), Western Lake Survey -  Phase I.
      The alkalinity map class from which each lake was selected is
      shown ( D: < 100 //eq L"1; A: 100-199 //eq L"1; X: 200-400
      //eq L"1; Omernik and Griffith 1986)	113

6-30  Relationship between ANC ( <  1000 //eq L"1) and elevation (m)
      for the Central Rockies (4D), Western Lake Survey - Phase I.
      The alkalinity map class from which each lake was selected is
      shown ( D: < 100 //eq L"1; A: 100-199 //eq L"1; X: 200-400
      //eq L~1; Omernik and Griffith 1986)	114

6-31  Relationship between ANC ( <  1000 //eq L~1) and elevation (m)
      for the Southern Rockies (4E), Western Lake Survey - Phase I.
      The alkalinity map class from which each lake was selected is
      shown ( D: < 100 //eq L"1; A: 100-199 /ueq L~1; X: 200-400
      //eq L"1; Omernik and Griffith 1986)	114

6-32  Relationship between ANC ( < 500 //eq L"1) and elevation (m)
      for lakes on the western and eastern slopes of the Sierra
      Nevada in California (4A) and Oregon Cascades in the Pacific
      Northwest (4B), Western Lake Survey - Phase I	116

  7-1  Locations of sampled lakes within geomorphic units in California
      (4A), Western Lake Survey - Phase I. Lakes located outside the
      boundaries were not included in estimates for these
      subpopulations	119
                                 XVIII

-------
                        Figures (continued)


Number                                                          Page

7-2  Locations of sampled lakes within geomorphic units in the Pacific
     Northwest (4B), Western Lake Survey - Phase I. Lakes located
     outside the boundaries were not included in estimates for these
     subpopulations	123

7-3  Locations of sampled lakes within geomorphic units in the
     Northern Rockies (40), Western Lake Survey - Phase I. Lakes
     located outside the boundaries were not included in estimates
     for these subpopulations	128

7-4  Locations of sampled lakes within geomorphic units in the Central
     Rockies (4D), Western Lake Survey - Phase I. Lakes located
     outside the boundaries were not included in estimates for these
     subpopulations	132

7-5  Locations of lakes within geomorphic  units in the Southern
     Rockies (4E), Western Lake Survey - Phase I. Lakes located
     outside the boundaries were not included in estimates for
     these subpopulations  	136

7-6  Trilinear diagram showing the relative abundance of major
     anions and cations for lakes in the largest geomorphic unit in
     California (Sierra Nevada), the Pacific  Northwest (Oregon
     Cascades), the Northern Rockies (Bitterroot Mountains), the
     Central Rockies (Wind River Range), and the Southern Rockies
     (Front Range), Western Lake Survey - Phase I. Ratios  are
     expressed as percent of total ionic concentration, increasing
     from 0 to 100 percent along the axes in the direction of the
     arrows	139

8-1  Cumulative frequency distributions [F(x>] for ANC (/ueq L"1) and pH
     for the Northeast (	), the Upper Midwest (	), the
     Southern Blue Ridge (	), and Florida (	), Eastern Lake
     Survey- Phase I (Linthurst et al. 1986), and the West (	),
     Western Lake Survey - Phase I	141

8-2  Concentrations of ANC (jueq L~1), pH, and concentrations of
     sulfate (/yeq  L~1) for lakes in the Northeast (Region 1),  Upper
     Midwest (Region 2), Southern Blue  Ridge (Subregion  3A), and
     Florida (Subregion 3B), Eastern Lake Survey - Phase I (Linthurst
     et al. 1986), and West (Region 4), Western  Lake Survey -
     Phase I. First quintile (61, 20th percentile), median (M, 50th
     percentile), and fourth quintile (Q4, 80th percentile) values are
     shown (O	1	D)	142
             Qi   M   0.4

8-3  Population estimates for the percent of lakes with concentrations
     of ANC < 0, < 50, and < 200 fjeq L'\  and pH < 6.0, for the
     Northeast, Upper Midwest,  Southern Blue Ridge, and Florida,
     Eastern Lake Survey - Phase I (Linthurst et al. 1986),  and West,
     Western Lake Survey - Phase I	143

8-4  Inverse cumulative frequency distributions [1 -F(x)] for sulfate
         L"1) and cumulative frequency distributions [F(x)] for calcium
         L"1)for the Northeast (	), Upper Midwest (	),

                                  xix

-------
                        Figures (continued)


Number                                                           Page

      Southern Blue Ridge (	), and Florida ( -	), Eastern
      Lake Survey - Phase I (Linthurst et al. 1986), and West (	),
      Western Lake Survey - Phase I	144

 8-5  Population estimates for the percent of lakes with concentrations
      of calcium < 50 /ueq L"1, sulfate > 50 fjeq L'\ DOC < 2 mg L"1,
      DOC > 6 mg L"1, and extractable aluminum > 50 //g L"1 for the
      Northeast, Upper Midwest, Southern Blue Ridge, and Florida,
      Eastern Lake Survey - Phase I (Linthurst et al. 1986) and West,
      Western Lake Survey - Phase I	144

 8-6  Concentrations of calcium (fjeq L"1), extractable aluminum
      (fjg L"1) in clearwater lakes,  and DOC (mg L"1) for lakes in the
      Northeast (Region 1), Upper Midwest (Region 2), Southern Blue
      Ridge (Subregion 3A), and Florida (Subregion 3B), Eastern Lake
      Survey - Phase I (Linthurst et al. 1986), and West (Region 4),
      Western Lake Survey - Phase I. First quintile (Q1f 20th
      percentile), median (M,  50th percentile), and fourth quintile
      (O.4, 80th percentile) values  are shown (O	1	D) 	145
                                          Qi  M  Q4

 8-7  Inverse cumulative frequency distributions [1 -F(x)] for extractable
      aluminum (//g L~1) in clearwater lakes, and cumulative frequency
      distributions [F(x)j for DOC (mg L"1) for the Northeast (	),
      Upper Midwest (	), Southern Blue Ridge (	), and
      Florida ( -	), Eastern Lake Survey - Phase I (Linthurst et al.
      1986) and West (	), Western Lake Survey -  Phase I	146

 8-8  Concentrations of ANC (fjeq L~1) and pH in selected
      subpopulations of lakes located in major geomorphic units in
      the Northeast, Upper Midwest, and Florida, Eastern Lake
      Survey - Phase I (Linthurst et al. 1986) and West, Western
      Lake Survey - Phase I. First quintile (Qi, 20th percentile),
      median (M, 50th percentile), and fourth quintile (Cu, 80th
      percentile) values are shown (O	1	D)	147
                                 Q!  M  Q<

 8-9  Concentrations of sulfate (/jeq L"1), calcium (/ueq L"1), and DOC
      (mg L"1) in selected subpopulations of lakes located in major
      geomorphic units in the Northeast,  Upper Midwest, and Florida,
      Eastern Lake Survey - Phase I (Linthurst et al. 1986) and West,
      Western Lake Survey - Phase I. First quintile (Q1f 20th
      percentile), median (M,  50th percentile), and fourth quintile
       Q4, 80th percentile) values are shown (O	1	D) 	148
                                          Q!  M  Q4

-------
                               Tables
Number                                                          Page

 2-1  Data Quality Objectives for Detectability, Precision, and Accuracy,
      Western Lake Survey - Phase I	4

 2-2  Non-target Lakes and Lakes Not Visited or Not Sampled, Western
      Lake Survey - Phase I	7

 2-3  Population Estimates of Median Lake Area and Median Elevation
      for Frozen Lakes and Sampled Lakes in Central Rockies (4D) and
      Southern Rockies (4E), Western Lake Survey - Phase I	8

 2-4  Forest Service Wilderness Areas and National Parks
      Containing Probability Sample Lakes, Grouped by Subpopulations
      Used in Data Analysis, Western Lake Survey - Phase 1	11

 2-5  Number of Lakes Shown on Small-scale (1:100,000) and
      Large-scale (1:62,500; 1:24,000) Maps, Western Lake
      Survey - Phase  I 	16

 2-6  Aliquots and Corresponding Measured Parameters, Western
      Lake Survey - Phase I 	22

 2-7  Maximum Holding Times Specified for Samples, Western Lake
      Survey - Phase  I 	22

 3-1  Number of Samples from Probability Sample Lakes Analyzed
      by the Analytical Laboratories, by Subregion, Western Lake
      Survey - Phase  I 	27

 3-2  Factors Used to Convert mg L~1 to /veq L~1 for Anion/Cation
      Balance Check,  Western Lake Survey - Phase I	30

 3-3  Factors Used to Convert mg L"1 to //S cm"1 for Determining
      Calculated Conductance, Western Lake Survey - Phase I  	30

 4-1  Detectability Based on Evaluation of Field Blank Data, Western
      Lake Survey - Phase I 	32

 4-2  Overall Among-Batch Precision Estimates Calculated from Field
      Synthetic Audit Samples, Western Lake Survey - Phase I	34

 4-3  Overall Among-Batch Precision Estimates Calculated from Field
      Natural Audit Samples, Western Lake Survey -  Phase I	35

 4-4  Overall Within-Batch Precision Estimates Calculated from Field
      Duplicates, Western Lake Survey - Phase I	36

 4-5  Overall  Field Laboratory Within-Batch Precision Estimates
      Calculated from Field Duplicate and Trailer Duplicate Data,
      Western Lake Survey - Phase I	37

-------
                         Tables (continued)


Number                                                         Page

 4-6  Overall Within-Batch Precision Estimates Calculated from
      Analytical Laboratory Routine/Duplicate Sample Data, Western
      Lake Survey - Phase I  	38

 4-7  Holding Time Samples Processed by Analytical Laboratories for
      the Calibration Study, Western Lake Survey - Phase I 	38

 4-8  Regression Statistics for the Difference Between Routine and
      Withheld Samples versus Holding Time by Laboratory, Western
      Lake Survey - Phase I  	39

 4-9  Results of Regression of Values for Primary Variables Measured
      on Samples Collected by Helicopter Crews (Dependent) Against
      Samples  Collected by Ground Crews for the Calibration Study
      Lakes, Western Lake Survey - Phase 1	44

4-10  Estimates of Bias between Analytical Laboratories Based on
      Synthetic and Natural Audit Samples (Audits), Paired Duplicate
      Samples  from the Calibration Study (Duplicates), and Split
      Samples  (Splits) with the Environmental Research Laboratory-
      Corvallis. The Model Results are Shown Using Lab 2 as trie
      Dependent Variable and Lab 1 as the Independent Variable	45

4-11  Comparison of Model Results of ANC on [Ca+2 + Mg+2] for the
      Central and Southern Rockies Using Data with Calcium Adjusted
      and Unadjusted for Possible Analytical Bias, Western Lake
      Survey -  Phase I	50

 5-1  Use of Weights in Combined Strata Estimation, Western Lake
      Survey -Phase I	54

 5-2  Composition of the  Alkalinity Map Classes: Numbers and
      Percentages of Lakes Having Measured ANC in Those Same
      Classes, Western Lake Survey - Phase I	55

 5-3  Population Statistics for ANC by Alkalinity Map Class, Western
      Lake Survey - Phase I  	56

 5-4  Population Statistics for ANC, Calcium, and  Sulfate in Shallow,
      Deep, Stratified, and Unstratified Lakes, Western Lake Survey -
      Phase 1	57

 5-5  Description of Target Population, Sample, and Weighting Factors
      Excluding Lakes > 2000 ha. Western Lake Survey - Phase 1	57

 5-6  Population Statistics for Physical Variables for Lakes in the
      Target Population, Western Lake Survey - Phase 1	58

 5-7  Population Estimates of the Percentage of Lakes in
      Categories by Lake  Type, Western Lake Survey -  Phase 1	58

 5-8  Population Estimates of the Percentage of Watersheds Dominated
      by Land Use/Land Cover Categories, Western Lake Survey -
      Phase 1	59

                                  xxii

-------
                         Tables (continued)


Number                                                          Page

 5-9  Population Estimates of Lakes with ANC < 50 or < 200 /ueq L"1,
      Western Lake Survey - Phase  I	60

5-10  Population Estimates and Population Statistics by  State for ANC
      and pH, Western Lake Survey - Phase I 	66

5-11  Population Estimates of Lakes with pH < 6.0, Western Lake
      Survey - Phase I 	66

5-12  Population Estimates of Lakes with Sulfate > 50 fjeq L"1, Western
      Lake Survey - Phase I  	66

5-13  Population Estimates of Lakes with Calcium < 50 //eq L~1, Western
      Lake Survey - Phase I  	67

5-14  Population Estimates of Clearwater Lakes with Extractable
      Aluminum > 50 //g L~1, Western Lake Survey - Phase I	67

5-15  Population Estimates of Lakes with Dissolved Organic Carbon
      < 2 mg  L"1 or > 6 mg L"1, Western Lake Survey - Phase I	68

5-16  Population Statistics for the Primary Variables, Western Lake
      Survey - Phase I 	71

5-17  Population Statistics for Secondary Variables (Nitrate, Ammonium,
      and Total Phosphorus), Western Lake Survey -  Phase I	71

5-18  Population Statistics for Secondary Variables (True Color,
      Turbidity, and Secchi Disk Transparency), Western Lake
      Survey - Phase I 	72

5-19  Population Statistics for Secondary Variables (Sodium, Potassium,
      and Magnesium), Western Lake Survey - Phase I 	72

5-20  Population Statistics for Secondary Variables (Iron, Manganese,
      and Total Aluminum), Western Lake Survey - Phase I 	73

5-21  Population Statistics for Secondary Variables (Silica, Dissolved
      Inorganic Carbon, Chloride, Conductance, and Bicarbonate),
      Western Lake Survey - Phase  I	73

5-22  Population Statistics for ANC,  Calcium, and Sulfate in Wilderness
      Area Lakes and Non-Wilderness Area Lakes, Western Lake
      Survey - Phase I 	74

5-23  Sample Statistics for Physical  and Chemical Variables for Special
      Interest Lakes, Western Lake Survey - Phase I	78

5-24  Population Statistics for Chemical and Physical Characteristics
      of Dilute Lakes, Western Lake Survey - Phase I	84

 6-1  Weighted Regression Statistics Comparing pH Measurements for
      Lakes Sampled Using Helicopters, Western Lake Survey -
      Phase I	85

-------
                        Tables (continued)


Number                                                         Page

 6-2  Regression Statistics for Calcium (< 500 /ueq L~1, Dependent)
      versus Magnesium ( <250/ueq L~1) by Subregion and Stratum,
      Western Lake Survey - Phase I	89

 6-3  Regression Statistics for Calcium (< 500 /ueq L"1, Dependent)
      versus Sulfate (< 100/ueq L~1) by Subregion and Stratum,
      Western Lake Survey - Phase I	89

 6-4  Regression Statistics for ANC (<200 /ueq L~1, Dependent) versus
      the Sum of Calcium and Magnesium (< 200 /ueq L"1) by
      Subregion and Stratum, Western Lake Survey - Phase I	93

 6-5  Relative Abundance of Major Cations and Anions Within
      Subregions Based on Population Estimates of Concentrations at
      the First Quintile (Qi) and Median (M) Values,
      Western Lake Survey - Phase I	95

 6-6  Regression Statistics for DOC (< 10 mg L"1, Dependent) versus
      True Color (< 50 PCU) by Subregion and Stratum, Western Lake
      Survey - Phase I	101

 6-7  Regression Statistics for the Sum of Major Anions (< 2000 /ueq L~1,
      Dependent) versus the Sum of Major Cations (< 2000 //eq L"1)
      by Subregion and Stratum, Western Lake Survey - Phase I	103

 6-8  Population Statistics for Anion Deficit, Western Lake Survey -
      Phase I 	103

 6-9  Regression Statistics for DOC (< 10 mg L"1, Dependent) versus
      Anion Deficit (< 100 /ueq L"1) by Subregion and Stratum,
      Western Lake Survey - Phase I	104

6-10  Spearman Correlation Coefficients (r) for Lake Area versus ANC,
      Calcium, and Sulfate, by Subregion and Stratum, Western Lake
      Survey - Phase I	108

6-11  Weighted Correlation Coefficients for ANC versus Lake Elevation
      by Geomorphic Unit, Western Lake Survey - Phase I	115

6-12  Population Statistics for ANC, Calcium, and Sulfate in High
      (> 3000 m) and Low (<3000 m) Elevation Lakes in California
      (4A), the Central Rockies (4D), and the Southern Rockies (4E),
      Western Lake Survey - Phase I	115

6-13  Population Statistics for ANC, Calcium, and Sulfate in Lakes on
      the Western and Eastern Slopes of the Sierra Nevada and
      Oregon Cascades, Western Lake Survey - Phase I 	116

6-14  Population Medians for Selected Parameters by Hydrologic Lake
      Type, Western Lake Survey - Phase I 	117

6-1 5  Estimated Hydraulic Residence Time (Years) for Drainage Lakes
      and Reservoirs by Subregion, Western Lake Survey - Phase I	117
                                 XXIV

-------
                         Tables (continued)
Number                                                          Page

 7-1  Population Estimates and Statistics for Primary Variables by
      Geomorphic Unit in California (4A), Western Lake Survey -
      Phase I  	120

 7-2  Population Statistics for Selected Variables for Lakes in
      Geomorphic Units in California (4A), Western Lake Survey -
      Phase I  	121

 7-3  Population Estimates and Statistics for Primary Variables by
      Geomorphic Unit in the Pacific Northwest (4B), Western Lake
      Survey - Phase I	125

 7-4  Population Statistics for Selected Variables for Lakes in
      Geomorphic Units in the Pacific Northwest (4B), Western Lake
      Survey - Phase I	1 25

 7-5  Population Estimates and Statistics for Primary Variables by
      Geomorphic Unit in the Northern Rockies (4C), Western Lake
      Survey - Phase I	130

 7-6  Population Statistics for Selected Variables for Lakes in
      Geomorphic Units in the Northern  Rockies (4C), Western Lake
      Survey - Phase I	130

 7-7  Population Estimates and Statistics for Primary Variables by
      Geomorphic Unit in the Central Rockies (4D), Western  Lake
      Survey - Phase I	133

 7-8  Population Statistics for Selected Variables for Lakes in
      Geomorphic Units in the Central Rockies (4D), Western Lake
      Survey - Phase I	133

 7-9  Population Estimates and Statistics for Primary Variables by
      Geomorphic Unit in the Southern Rockies (4E), Western Lake
      Survey - Phase I	1 37

7-10  Population Statistics for Selected Variables for Lakes in
      Geomorphic Units in the Southern Rockies (4E), Western Lake
      Survey - Phase I	137

 8-1  Physical Characteristics of Lakes Sampled During Phase I of the
      Eastern and Western Lake Surveys	141
                                  XXV

-------
                       Related Documents^
Anonymous. 1984. National Surface Water Survey, National Lake Survey -
Phase I, Research Plan. U.S. Environmental Protection Agency, Washington,
D.C. (internal document).

Bonoff, M.B. and A.W. Groeger. 1987. National Surface Water Survey, Western
Lake Survey -  Phase I, Field Operations Report.  Internal Report. EPA/600/
8-87/018. Environmental Monitoring Systems Laboratory, U.S. Environmental
Protection Agency, Las Vegas, Nevada. 26 pp.

Eilers, J.M., D.J. Blick, and M.S. DeHaan. 1987. National Surface Water Survey,
Western Lake Survey - Phase I, Validation of the Western Lake Survey - Phase
I  Data Base.  Environmental  Research  Laboratory,  U.S.  Environmental
Protection Agency, Corvallis, Oregon.

Kanciruk, P., M.J.  Gentry, R.A. McCord, L.A. Hook, J.M. Eilers, and M.D. Best.
1987. National Surface Water Survey, Western Lake Survey  - Phase I, Data
Base Dictionary.  Oak Ridge National Laboratory, Technical Manual ORNL/
TM-10307, Oak Ridge, Tennessee.

Kerfoot, H.B. and  M.L. Faber. 1987. National Surface Water Survey, Western
Lake Survey  - Phase I, Analytical Methods  Manual.  EPA/600/8-87/038.
Environmental Monitoring Systems Laboratory, U.S. Environmental Protection
Agency, Las Vegas, Nevada.

Silverstein, M.E., S.K. Drous3, J.L. Engels, M.L. Faber, andT.E. Mitchell-Hall.
1987a. National Surface Water Survey, Western Lake Survey -  Phase I, Quality
Assurance  Plan.  EPA/600/8-87/026.  Environmental  Monitoring  Systems
Laboratory, U.S. Environmental Protection Agency, Las Vegas, Nevada.

Silverstein, M.E.,  M.L. Faber,  S.K. Drouse, and T.E.  Mitchell-Hall. 1987b.
National  Surface  Water Survey, Western Lake Survey - Phase  I, Quality
Assurance Report.  Environmental Monitoring  Systems  Laboratory, U.S.
Environmental Protection Agency, Las Vegas, Nevada.
'Many of the documents are in draft form at the time of this publication


                                 XX vi

-------
                        A cknowledgments
The Western Lake Survey built upon the approach used in the Eastern Lake
Survey and benefited immensely from the experience gained in the previous
effort. Consequently, although many of the personnel involved in the Western
effort differed, a deep debt of gratitude is owed to all individuals involved
in the design, administration and execution of the Eastern Lake Survey. The
Western  Lake Survey  represented a  challenge  both  administratively and
logistically. This  challenge was met through  the  dedication and energy of
hundreds of individuals coordinated through the U. S. Environmental Protection
Agency and the  USDA Forest Service.  The short time period  identified for
sampling in  the West, the need to access nearly two-thirds of the lakes on
foot, and the rigors of sampling remote, high elevation lakes  in pre-winter
conditions made the Western Lake Survey a great  challenge. The authors
can acknowledge only a few of those who performed key roles  in the Survey
but readily recognize the many others, without whom the Western Lake Survey
could not have succeeded.

L. Edwin Coate,  then the Deputy  Regional Administrator of EPA Region X,
was the inspirational and dedicated project leader who provided the patience,
experience, and vision necessary to guide this project through many uncertain
times. He also created a working management structure combining elements
from EPA  Region X, EPA Office of Research and  Development Laboratories,
and the Forest Service that allowed personnel to work together to complete
the Survey successfully. Without his leadership and belief in the critical need
for data on western lakes, this study may never have been completed.

Courtney Riordan, Director of the EPA Office of Acid Deposition, Environmental
Monitoring and Quality Assurance, Office of Research and Development, was
instrumental in  early  decisions to perform the Western  Lake Survey and
supported  the entire effort. Gary Foley, Director of the EPA Acid Deposition
and Atmospheric Research Division worked to insure that the  Survey would
progress from the design stage to the implementation phase.

The close cooperation between the U.S. Environmental Protection Agency and
the Forest Service could not have been realized without the support and
encouragement of Lee Thomas, Administrator  of the U.S. Environmental
Protection Agency and Maxwell Peterson, Chief of the Forest Service. Their
joint leadership and commitment to the successful completion of this regional
assessment are gratefully acknowledged.

Logistics, training and coordination were provided by  several  organizations
and many  dedicated individuals. The following list identifies the  major activity
area and  lists the  individuals involved and their affiliations (abbreviations
defined at  end of this section) at the time of the Survey:
                                 XXVII

-------
                        Logistics Coordination
           Lee Marshall
           Rick Ross
           Richard Dearsley
           Robert Cusimano
           Stephen Pierett
           Kenneth Asbury
           John Baker
                 EPA, Region X
                 FS
                 FS
                 Northrop Services, Inc.
                 Lockheed-EMSCO
                 Lockheed-EMSCO
                 Lockheed-EMSCO
                       Base Station Operation
Carson City, Nevada:

Arnold Den
Robert Doty
Kenneth Greenberg

Wenatchee, Washington:

David Tetta
Carl Anderson
Ray Peterson

Missoula, Montana:

Rick Claggett
Michael Goggin
Douglas Fox

Bozeman, Montana:

C. E. "Cornie" Runas
Peter Stender
William Warner

Aspen. Colorado:

Lester Sprenger
Dennis Haddow
Larry Svoboda
               EPA, Region X, Base Coordinator
               FS, Field Manager
               EPA, Region X, Duty Officer
               EPA, Region X, Base Coordinator
               FS, Field Manager
               EPA, Region X, Duty Officer
               EPA, Region VIII, Base Coordinator
               FS, Field Manager
               EPA, Region VIII, Duty Officer
               EPA, Region VIM, Base Coordinator
               FS, Field Manager
               EPA, Region VIII, Duty Officer
               EPA, Region VIII, Base Coordinator
               FS, Field Manager
               EPA, Region VIII, Duty Officer
Ken Asbury
Kevin Cabbie
John Henshaw
Franklin Morris
                      Training and Procurement

                    Lockheed-EMSCO (EMSL-LV):
   John Baker
   Linda Drewes
   Cindy Mayer
   David Peck
Barry Baldigo
Cindy Hagley
Molly Morison
James Wilson
Kevin Cabbie
Mel Knapp
Linda Drewes
C. Hunter Nolen
                     Field Laboratory Coordinators

                    Lockheed-EMSCO (EMSL-LV):
   Gerald Filbin
   Ky Ostergaard
Alan Groeger
 Field Laboratory Supervisors

Lockheed-EMSCO (EMSL-LV):

   John Henshaw          Molly Morison
   David Peck
                                XXVIII

-------
Michael Bonoff
Cindy Hagley
James Wilson
                          Helicopter Crews

                    Lockheed-EMSCO (EMSL-LV):
      Karen Cougan
      Cindy Mayer
Barry Gall
Mark Mitch
Allen Belenz
Betsy Dickes
Lesa Madison
Peggy Permann
Brad league
                      Field Laboratory Analysts

                    Lockheed-EMSCO (EMSL-LV):
      James Crawford
      Elizabeth Hill
      Marty McKaig
      Susan Piacentine
      Gary Turner
David Grouse
David Janick
Katy Nedrow
Elizabeth Sweeney
Sandra Watford
Kathleen Cahall
Julie Segars
      Ground Crews-Logistics

   Lockheed-EMSCO (EMSL-LV):

      Robert Danehy          Thomas Loranger
      James Williams
Steve Morris
Communications Center in Las Vegas

   Lockheed-EMSCO (EMSL-LV):

      Valerie Sheppe         Brenda Whitfield
Byron Blasdell
Charles Shoemaker
        Engineering Support

   Lockheed-EMSCO (EMSL-LV):

      Jeff Love               Michael Reese
      John Wengert
Daniel Allison
Bryant Hess
John Lau
Maria Maxwell
                    Technical and Clerical Support

                    Lockheed-EMSCO (EMSL-LV):
      Ramon Denby
      Daniel Hillman
      Linda Marks
      Peggy Oakes
Annalisa Hall
Henry Kerfoot
Richard Maul
Kathleen Hurley
Sharon Ziminski
  Northrop Services, Inc. (ERL-C):

      Regina O'Brien         Roze Vanden Berg
John Hazard
        Statistical Support

          Forest Service:

      Ronald McRoberts
Sandy Verry
                   Lockheed-EMSCO (EMSL-LV):

                         Martin Stapanian
                                XXIX

-------
                      Montana State University:

                           Daniel Goodman
James Blick
  Northrop Services, Inc. (ERL-C):

                             Mark DeHaan
Thomas Permutt
     Systems Applications, Inc.

                             Alison Pollack
Brooke Abruzzesse
Andrew Kinney
        Geographic Support

   Northrop Services, Inc. (ERL-C):
      Sharon Clarke          Colleen Johnson
      Barbara Rosenbaum
Wesley Kinney
Quality Assurance and Quality Control

          EPA, EMSL-LV:

                             Robert Schonbrod
                    Lockheed-EMSCO (EMSL-LV):
Karen Cougan
Marianne Faber
Tuijauna Mitchell-Hall
      Lynn Creelman
      Carol MacLeod
      David Peck
Sevda Drouse
Dean Mericas
Sharon Chandler
Cindy Wear
  Data Management and Data Entry

              ORNL:

      Paul Kanciruk           Sheila Ladd
Merilyn Gentry
           SAIC (ORNL):

      Les Hook
Raymond McCord
Stephen Norton
Frank Sanders
John Turk
Robert Wetzel
Robert Wissmar
          Report Reviewers

          University of Maine, Orono, Maine
          University of Wyoming, Laramie, Wyoming
          U.S. Geological Survey, Denver, Colorado
          University of Michigan, Ann Arbor, Michigan
          University of Washington, Seattle, Washington
The Office of Aircraft Services, U.S. Department of the Interior, arranged for
helicopter support and provided flight training  for all helicopter  samplers.
Training for  helicopter and ground crews was coordinated and performed by
John Baker of Lockheed-EMSCO (EMSL-LV). Kevin Cabbie, Lockheed-EMSCO,
was instrumental in selecting the locations of the base stations and ensuring
that required services were provided to the field laboratory and logistics room
at each base site.
                                xxx

-------
Several universities were involved in contributing to the Western Lake Survey:
Oregon State  University-Corvallis,  Oregon; State University of New York-
Oswego,  New York; and Western Washington  University-Bellingham,
Washington. State agencies also were important in  providing information on
lakes in their respective jurisdictions and in providing special assistance in
acquiring permission for access. All are acknowledged for their contribution
to the survey.

Many individuals were provided by the Forest Service to coordinate and perform
the  ground sampling of lakes in  wilderness areas. The enthusiasm,
professionalism, and dedication  of these people were a key component of
the  success  of this  project and  their stellar  performance  is gratefully
acknowledged.

Finally, this report was typed by Roze Royce of Northrop (ERL-C). This vital
task was completed accurately, timely, and cheerfully, and the authors would
like to express their  gratitude for her important contribution to completing
this report. Regina O'Brien coordinated preparation of figures, and the graphic
artists were Linda Haygarth and Kris Daniels. The assistance of Avis Newell
of Northrop (ERL-C) and Thomas Loranger of Western Washington University
in finalizing the report is equally appreciated.
Abbreviation
                  Definition
EPA

FS


Lockheed-
EMSCO

ERL-C


EMSL-LV



ORNL

SAIC

SAI
United States Environmental Protection Agency

United States Department of Agriculture-Forest
Service

Lockheed Engineering and Management
Services Company, Incorporated

U.S. EPA Environmental Research Laboratory at
Corvallis, Oregon

U.S. EPA Environmental Monitoring Systems
Laboratory at
Las Vegas, Nevada

Oak Ridge National Laboratory

Science Applications International Corporation

Systems Applications, Inc.

-------
                                     Executive Summary
Background

Phase  I of the  Western Lake Survey (WLS-I) was
conducted in the fall of 1985 by the U.S. Environ-
mental  Protection Agency (EPA) as  part  of  the
National Surface  Water Survey (NSWS). The first
phase of the NSWS was designed to determine the
present chemical status of surface waters in regions
of the United States that contain the majority of lakes
that are considered to be at risk as a result of acidic
deposition. The  WLS-I was conducted as part of the
National Acid  Precipitation  Assessment Program
(NAPAP). It contributes directly to one of NAPAP's
principal objectives: the quantification of the extent,
location, and characteristics of sensitive  and acidic
lakes and streams in the United States.

The WLS-I  was  conducted  in  five areas  in  the
western United  States (see inside cover). The areas
surveyed included the major mountain ranges in the
West, containing  the Sierra Nevada, the Cascade
Range, and the Rocky Mountains. Five subregions,
delineated  on the basis of  similar  physiographic
characteristics,  were identified: California (4A), the
Pacific Northwest (4B), the Northern Rockies (4C),
the Central Rockies (4D), and the Southern Rockies
(4E). An analogous survey of eleven subregions was
conducted  in the  fall of 1984 in three regions of
the eastern United States. These regions were the
Northeast, the Upper Midwest, and the Southeast.

The primary objectives of the WLS-I  were  to
determine  in potentially  sensitive  areas  of  the
western United States: (1) the percentage (by number
and area) and location  of lakes that are acidic; (2)
the percentage  (by number and area) and location
of lakes  that have  low acid neutralizing capacity
(ANC);  and (3) the chemical characteristics  of lake
populations, providing a data base for selecting lakes
for future studies.

Because not all lakes in the western United States
could be sampled,  a  statistical procedure was
developed  for  selecting a  subset  of  lakes as a
probability  sample.  Within  each subregion, three
areas expected to  be  dominated  by lakes with
alkalinity < 100 /ueq L~1, 100-199 /ueq  L"1, or 200-
400 //eq  L~1 were identified on maps. A systematic
procedure with  a  random starting point  was used
to select lakes  from  each of the three areas. This
design ensured that each lake within each of the
three alkalinity map classes was equally likely to be
selected; thus, the  physical and chemical charac-
teristics of the population of lakes could be estimated
from  sample results  with a known  degree of
confidence. These lakes are referred to as probability
sample lakes. The number of lakes in the probability
sample upon which population estimates were based
was 719, representing an estimated 10,393 lakes
in the target population.

To  maximize the number of lakes that  could  be
surveyed, a single sample was collected from each
lake.  Fall was selected as the optimal period for
sampling because chemical variability within a single
lake is expected to be low in most temperate lakes
as a result of circulation of the water column. Data
from  the small  percentage of  lakes  that were
stratified did not alter the results. The results of the
WLS-I are applicable to the fall season of sampling;
representation of other periods during the year in
western  lakes is unknown. For example,  nitrate
concentrations in the lakes from the fall survey may
be much lower than values measured in the spring.

The prescribed method of sample collection  for the
lake survey within the NSWS was by helicopter. Of
the 752 lakes identified for sampling in WLS-I, 455
were located in wilderness areas, to which motorized
access is not permitted. Thus, lakes located outside
wilderness areas were accessed by hel icopter crews,
and wilderness  lakes  were accessed  by ground
crews. Ground sampling was conducted by person-
nel from the U.S. Department of Agriculture-Forest
Service. To evaluate whether there was a difference
between data collected by ground crews and  by
helicopter crews,  the Forest  Service  granted
permission to conduct a special study in which 45
lakes  located  in wilderness areas were sampled
using  both  methods. The  results  of  this  study
indicated that samples collected  using the ground
method, even with longer holding times, yielded data
that were  comparable to  those  obtained from
samples collected using helicopters.

Water samples  were delivered  to  mobile field
laboratories for preliminary processing before they
were sent to one of two analytical laboratories for
chemical analysis. Extensive quality assurance and
                                             XXXIII

-------
data management programs were implemented to
ensure that all sample collection activities, analytical
methods,  and  data  reporting  procedures  were
performed  within  the limits  of pre-established
criteria designed to meet specifically designed data
quality objectives.  For  nearly all variables,  these
criteria were met; however, statistically significant
laboratory biases were observed for some variables.
The small magnitude of the bias had no perceptible
effect on the population estimates, but users of  the
data should be  cognizant of possible limitations in
the data.

This report, Chemical Characteristics of Lakes  in  the
Western United States, consists of two volumes.
Volume I,  Population  Descriptions and  Physico-
chemical Relationships, provides details  about  the
design and  implementation of the WLS-I, discusses
results,  and presents conclusions based on  these
results.  Volume II,  Data Compendium for Selected
Physical and Chemical Variables, contains additional
estimations of the characteristics of the populations
of lakes sampled  within each  area, descriptive
statistics for these lake  populations, and a data
compendium of site characteristics and chemical
variables for each lake sampled.

The purpose of  this report is to describe, within  the
context  of the survey design, the chemical charac-
teristics of  lakes in the western United States,  the
sampling and analytical protocols,  and the quality
of the data obtained. These aspects of the WLS-I,
described in detail in Sections 2 and 3, must be
understood before conclusions beyond the scope of
the original objectives can be drawn. The WLS-I data
cannot  be  used alone  to  establish a relationship
between the present chemical status of lakes  and
the occurrence of acidic  deposition.  However,  the
data can be used to determine associations among
variables. The data base also can be used to  refine
our understanding of the response of lakes to  acidic
deposition by permitting subpopulations of lakes to
be  selected to  test hypotheses  regarding aquatic
response to acidic deposition.

Selected  Results
To meet the three primary objectives of the survey,
the data from the WLS-I were used to examine  the
characteristics of lakes in the West, both within and
among  the  five western  subregions.  Population
estimates were calculated for the target population
of lakes (10,393 lakes; > 1 ha and  < 2000 ha) in
the West in the five  subregions  sampled. The
population estimates apply only to those geographic
areas identified for the study (see inside cover).
Estimates were also calculated for special popula-
tions of lakes including those located within various
geomorphic units generally coinciding with moun-
tain ranges, dilute lakes (those with conductance
< 10 /jS cm"1), and lakes within wilderness  areas
and  national parks. Associations among  variables
also  were examined through correlative analyses.

Extent and Location of Lakes with Low Acid
Neutralizing Capacity (ANC) and Low pH

Of the probability sample lakes, only one was acidic
(ANC < 0 Aieq  L~1). Only one lake had pH < 5.0.
An estimated 16.8 percent of the target population
of lakes in the western United States had low ANC
(< 50 £ 6.0.

California (Subregion 4A) had the highest estimated
number and percentage of  lakes with low ANC in
the  West.  Lakes  with low  ANC  were located
throughout  the  sampling  area in  California,  but
particularly in the Sierra Nevada.
    Pacific Northwest (Subregion 4B) had the second
highest estimated  number  and percentage of  low
ANC lakes in the West. These lakes were located
along the Cascade crest in Oregon and Washington.

The Northern Rockies (Subregion 4C) had the third
highest estimated  number of lakes with low ANC.
Most  of these lakes were located in  the Bitterroot
Range along the Idaho-Montana  border and in the
Sawtooth  Mountains of Idaho.

The Central Rockies (Subregion 4D) had the fourth
highest percentage of lakes with low ANC. Low ANC
lakes in this Subregion were generally located in the
Wind  River Range and the  Big Horn  Mountains in
Wyoming,  the  Absaroka-Beartooth Mountains in
Wyoming  and Montana,  and the Uinta Mountains
in Utah.

The Southern Rockies (Subregion 4E) had the fewest
lakes  with low ANC. These were located  primarily
in the San Juan  Mountains and the Front Range
in Colorado.

Special Populations

Many  of  the lakes  in the West had very  low
concentrations of dissolved constituents. Concentra-
tions  of most measured variables, such as base
cations (Ca+2, Mg+2, K+, Na+), major anions (HC03~,
SO/2, Cl~) and  nutrients (total  P, N03~,  NH4+) were
very low in these dilute lakes (i.e., conductance <
10 /uS cm"1).  Western lakes were nearly all  clear-
water (true  color  <  30  PCU),  and  were  very
transparent.

Lakes in the wilderness areas, representing approx-
imately two-thirds  of the  total target  population in
the West,  had substantially lower concentrations of
major  anions  and cations than  lakes  in non-
wilderness areas. Median ANC values for lakes in
the wilderness areas were  91.4 peq  L"1 compared

-------
Table E-1. Population Estimates and Statistics for
Phase 1
Subregion
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West
Subregion
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West
n*** number of
ft estimated
PC estimated
ftc estimated

(4A)
(4B)
(4C)
(4D)
(4E)
(4)

(4A)
(4B)
(4C)
(4D)
(4E)
(4)
n***
149
159
143
129
139
719
n***
149
159
143
129
139
719
ft
2401
1706
2379
2299
1609
10393
ft
2401
1706
2379
2299
1609
10393
Acid Neutralizing Capacity and pH by Subregion, Western Lake Survey -

PC
0367
0195
0.127
0.069
0046
0.168

PC
0.013
0024
0000
0.013
0000
0010
lakes sampled.
target population size.
proportion of lakes with given criteria
number of lakes with given criteria.
ANC< 50A 50 jueq L"1 and with  calcium < 50 /ueq L"1. Of
the few clearwater western lakes (16) with extrac-
table aluminum > 50 /ug L"1, most were located in
this subregion.

As  was  observed for lakes in other subregions,
concentrations of nitrate, ammonium  and  total
phosphorus were very  low. The  highest median
values for silica, manganese, and  chloride were
observed in the  Pacific  Northwest. Concentrations
of other chemical constituents were similar to those
measured in the three Rocky Mountain subregions.

Northern Rockies (4C)--

In the Northern Rockies, an estimated 23.7 percent
of the target population of lakes had conductance
values < 10 //S  cm"1, similar to the percentage of
dilute lakes estimated for the Pacific Northwest, but
much higher than the percentages estimated for the
Central and  Southern  Rockies.  Relative to  other
western  subregions, lakes in the Northern Rockies
were intermediate in  estimated  median  sulfate
concentration and percentage of lakes with sulfate
> 50 /ueq  L"1. The median values for calcium and
extractable aluminum were the second  highest for
any subregion surveyed,  although all  clearwater
lakes sampled had extractable aluminum <50//g L"1.
Dissolved organic carbon  concentrations were also
low; only an estimated 8.2 percent of the lakes had
                                               XXXV

-------
DOC > 6 mg L"1. Median values for most secondary
variables  were intermediate relative  to  those
estimated for other subregions.

Central Rockies (4D)~

It was estimated that slightly more than  10 percent
of the lakes in the Central Rockies were dilute, which
was substantially  less  than  the percentages
observed in  California, the Pacific Northwest, and
the Northern Rockies. This subregion hadthe second
highest estimated median sulfate  value, but the
lowest percentage and number of lakes in the West
with sulfate > 50 //eq L"1.  Lakes in the Central
Rockies had the lowest median calcium  concentra-
tion of the three  Rocky Mountain subregions and
the second lowest number and percentage of lakes
in the West with calcium < 50 /ueq L"1. The estimated
median concentration of extractable aluminum for
clearwater lakes was the lowest among all  subre-
gions,   and  no lakes were estimated to  have
concentrations of extractable  aluminum > 50 /ug L"1.
As  in  other western  subregions,  lakes  in  this
subregion  had generally low DOC concentrations.

Southern Rockies (4E)~

Of the five western subregions, it was estimated-that
the Southern Rockies had the fewest dilute lakes
and the highest median calcium concentrations. An
estimated 33.7 percent of the lakes (543)  had sulfate
concentrations  > 50 /ueq  L"1,  nearly  twice the
percentage  and number  of  lakes in the  Pacific
Northwest. The lakes with high  sulfate  concentra-
tions were generally associated with high ANC lakes.
Calcium concentrations exceeded 50/ueq L"1 in 96.6
percent of the lakes. All clearwater lakes sampled
had extractable aluminum  concentrations  < 50
/ug  L"1. Dissolved organic  carbon was generally
higher than in the other subregions.

Lakes  in the Southern Rockies  had the highest
estimated median values for nitrate, total phospho-
rus, sodium, potassium, magnesium, iron, dissolved
inorganic  carbon, conductance,  and  bicarbonate.
These lakes also had  the highest median true color
and turbidity,  resulting  in  the  lowest median
transparency of any western subregion.

Associations Among  Variables

In most western lakes surveyed, the dominant cation
was calcium and the dominant anion  was bicarbo-
nate. Sodium was the second most abundant cation
in California and the Pacific Northwest, probably due
to mineral weathering in California and the Oregon
Cascades and to mineral weathering and atmospher-
ically deposited sea  salt in the North  Cascades. In
the Rocky  Mountains, magnesium was the second
most abundant cation, except in lakes with lower
total cation concentrations in the Northern Rockies,
where sodium was approximately equal to magne-
sium.  Based  on  the  relative abundance of major
anions and cations and on patterns observed in ratios
of major  cations and  major  anions,  most lakes
surveyed in the West can be classified as calcium
bicarbonate systems.

A positive relationship between calcium and sulfate
concentrations was observed  in the  lakes of the
Rocky Mountains, suggesting that some sulfate  is
derived  from  mineral sources.  An analysis of the
relationship between  ANC and  the sum of calcium
and magnesium in the western subregions, for which
the intercepts of the regressions were near zero and
the slopes were near  one, suggests that wide-scale
acidification of lakes has not occurred in these areas.

The relationship between low ANC in lakes  and
bedrock type indicates that geological characteristics
of watersheds in the  western United States play a
major role in determining the ANC  levels in surface
waters.  Consequently, bedrock type is an important
factor determining the potential sensitivity, in terms
of static ANC  measurements, of western lakes to
acidic deposition.

No relationship was found between  lake area  and
water chemistry for lakes < 30 ha, except that lakes
<  4 ha  in the Northern Rockies were  more dilute
than those lakes > 4 ha. Lakes < 30 ha had lower
ANC values than lakes  >  30 ha. Elevation of lakes
was not always correlated with lake water chemistry.
Acid neutralizing capacity decreased with elevation
in the Wind River Range in Wyoming and in the
California  subregion,  but  no relationship was
observed  in the Pacific Northwest, the Northern
Rockies, or the Southern Rockies.

Differences in water chemistry among hydrologic
lake types were observed. Reservoirs in  California
and the Southern Rockies had  substantially higher
median  ANC and calcium  values than did drainage
or seepage lakes; in  California  and  the Pacific
Northwest, drainage lakes had  slightly higher
median ANC, calcium  and sulfate concentrations and
lower median DOC concentrations than did seepage
lakes. In  the  Rocky  Mountain subregion, similar
results were found for sulfate and DOC, but seepage
lakes  had somewhat  higher  ANC than did drainage
lakes.


Comparison of Western Lakes to Eastern
Lakes

In contrast to lakes surveyed in the eastern United
States,  no acidic (ANC < 0 peq  L~1) lakes were
sampled in the West  that  were not associated with
hot springs. Of the  four  regions  surveyed in the
                                              xxxvi

-------
National  Lake Survey, the highest  percentage  of
lakes with ANC < 200 //eq L"1 was observed in the
West. Only one  lake with pH < 5.0 was sampled
in the West;  in the East,  10 percent of the lakes
in the Adirondacks,  9.4 percent  in  the  Upper
Peninsula of Michigan, and 12.4 percent in Florida
were estimated to have pH < 5.0. Only 1.0 percent
of western lakes had pH < 6.0, whereas  in the East,
the percentage ranged from 9.6 percent in the Upper
Midwest to 32.7 percent in Florida.

Sulfate concentrations in  most western  lakes were
substantially lower than those in  most eastern lakes.
The  estimated median sulfate concentration in the
Northeast was approximately six times that for
western lakes as a whole. In the West, only  13.5
percent  of the lakes  had sulfate > 50 //eq L"1,
whereas 97.3 percent  of the lakes in the Northeast,
61.1  percent in the Upper Midwest, and 68.1 percent
in Florida had sulfate > 50 //eq L"1.

Estimated  regional  calcium  concentrations  in
western lakes were also substantially lower than in
eastern lakes. The total estimated number of lakes
in the western target population with calcium < 50
//eq  L"1  exceeded that  for all eastern regions
combined. The  number of clearwater lakes  in the
West with extractable aluminum concentrations
> 50 //g L"1 was much smaller than that observed
in the East. Dissolved organic carbon concentrations
were also  much  lower in western  lakes than  in
eastern  lakes with the exception of lakes  in the
Southern Blue Ridge subregion; 69.7 percent of the
lakes in the West were estimated to have DOC < 2
mg L~1, whereas only 8.1 percent of the lakes  in
the Northeast, 1.3 percent in the Upper Midwest,
and  11.5 percent in Florida had DOC < 2 mg L"1.

Conclusions
Within  the areas of  the  western United States
sampled and covered by the design of the Western
Lake Survey - Phase I,  it is concluded that:

• More than  one-fourth (26.6  percent) of the
   western  lakes are dilute  (conductance  < 10
   //S cm"1). The areas sampled in the western United
   States contain a much higher percentage of lakes
   with low concentrations of dissolved substances
   than do areas sampled in three regions (Northeast,
   Upper  Midwest, and Southeast) in the eastern
   United States. The lowest Qi value (20th percen-
   tile) for conductance measured in the East was
   20.7 pS  cm"1 and the  minimum value was 7.8
   fjS cm"1. In contrast, 20 percent of the lakes  in
   the West had conductance values < 8.3 //S cm"1
   and the minimum measured conductance in the
   West was 1.6//S cm"1.

• Calcium concentrations in the West varied widely
   among subregions;  56.5 percent of the lakes  in
  California had calcium <50//eq L 1, whereas only
  3.4 percent of the lakes in the Southern Rockies
  were below this criterion.  Overall,  the median
  concentration of calcium for lakes in the  West
  was low (92.4 //eq L~1) compared to the Northeast
  (177.4 //eq L"1), the Upper Midwest (238.2//eq L"1),
  and Florida (238.3 //eq L"1). Lakes in the Southern
  Blue Ridge had a comparable median value  for
  calcium (104.7 //eq L"1).

• Median sulfate concentrations  were  extremely
  lowthroughoutthe West, ranging from 6.6//eq L~1
  in California  to  34.6 //eq L"1  in the Southern
  Rockies.  In comparison, median concentrations
  for lakes in the East were 115.4 //eq L"1 for the
  Northeast, 57.1 //eq  L~1 for the Upper Midwest,
  31.8 //eq L"1  for the Southern  Blue  Ridge, and
  93.7 //eq L~1 for Florida.

• Extractable aluminum  concentrations in  clear-
  water lakes were very low throughout the West.
  Only 0.2 percent of  the lakes  in the  West had
  extractable aluminum concentrations exceeding
  50 //g L"1 compared  to 5.5 percent of the  lakes
  in the Northeast, and 7.4 percent of the lakes in
  Florida. These results are, again, similar to those
  for the  Southern Blue  Ridge,  where no  lakes
  exceeded this criterion.

• Concentrations of dissolved organic carbon were
  also low throughout the West; only 5.4 percent
  of the lakes in the West had DOC values exceeding
  6  mg L"1, whereas  this criterion  was exceeded
  for 26.4  percent of the lakes in the Northeast,
  62.9 percent in the Upper Midwest, and 68.9
  percent in Florida. Again, the Southern Blue Ridge
  shows a similarity  to  western  lakes with 6.1
  percent of the lakes with DOC > 6 mg L~1.

• pH values were  not  low in the  West,  where 99
  percent of lakes had values  greater  than 6.0.
  Median pH values for the subregions ranged from
  6.94 in California to 7.60 in the Southern Rockies.

• Of fhe subregions sampled in the West, California
  had  the  largest number (881)  and percentage
  (36.7%) of lakes with low ANC (< 50 //eq L"1);
  followed  by the Pacific Northwest  (333  lakes;
  19.5%). The Southern Rockies had the smallest
  number and lowest percentage of lakes with low
  ANC (74  lakes; 4.6 %).

• The  lakes in wilderness areas had  much lower
  concentrations of ANC with a  median value  of
  91.4 //eq L~1 compared to 282.7 //eq L"1 for non-
  wilderness lakes.

• The Sierra Nevada (the area which contained the
  majority  of lakes sampled  in California) had the
  largest estimated number of low ANC lakes (834)
                                             XXXVII

-------
and dilute lakes (1311) of any geomorphic unit
in the West.

No lakes sampled in the West had ANC
-------
                                          Section 1
                                         Introduction
Assessment of the effects of acidic deposition on
lakes  in the  western United States  has received
considerable  scientific attention relatively recently
(Lewis and Grant 1979; McColl 1981; Turk 1983;
Smith and Alexander 1983; Kling and Grant 1984;
Rothetal. 1985; Hidy and Young 1986). Few of these
studies reported decreases  in acid neutralizing
capacity or pH in lakes as a result of acidic deposition.
Many of the  lakes in the West have  low alkalinity
(Omernik and Griffith  1986),  and are,  therefore,
considered to  be potentially  sensitive to acidic
deposition; however, their  remote  location  and
difficult access have resulted  in a limited under-
standing  of  their chemical  characteristics.  In
addition, regional patterns of acidic deposition in the
West  have not been  easily identifiable  because of
an insufficient number of deposition monitoring sites
(Munger and Eisenreich  1983;  Stensland 1984;
Stensland et  al. 1986; Hidy and Young  1986).  This
lack of monitoring sites increases the difficulty of
assessing the extent to which western lakes are at
risk as a result of acidic deposition. In contrast, in
the eastern United States where levels of acidic
deposition have been monitored much more exten-
sively, the relationships between acidic  deposition
and surface water chemistry have been studied more
intensively (Galloway et al.  1976;  Likens 1976;
Altshuller and Linthurst 1984).

To  provide a  high  quality data base  for use in
quantifying the present chemical status of western
lakes,  the  U.S.  Environmental Protection  Agency
(EPA)  designed and implemented the Western Lake
Survey - Phase I (WLS-I) as part of the National
Surface Water Survey (NSWS). The NSWS is a three-
phase  project with two components, the National
Lake Survey (NLS) and the National Stream Survey.
Phase I is designed to quantify the present chemical
status of U.S. surface waters.  Phase  II is designed
to assess the chemical variability and to quantify
biological resources in  surface waters;  and Phase
III will identify temporal  trends in surface water
chemistry and biology.

Initially, the NLS was scheduled to be  conducted
simultaneously in four regions  of the  United States
(the  West,  Upper  Midwest,  Northeast,  and
Southeast)  during the  fall of  1984.  As planning
progressed, it  became  apparent that logistical
considerations for the western effort were compli-
cated by  rigorous  terrain and  weather  and  by
restrictions on motorized access to sampling sites.
Consequently, the NLS was divided into the Eastern
Lake Survey - Phase  I (ELS-I) during which 1798
lakes were sampled in the fall of 1984 (Linthurst
et al. 1986) and the WLS-I during which 752 lakes
were sampled in the fall of 1985. Deferring the WLS-I
permitted  the special conditions in the West to  be
considered fully before field activities were imple-
mented.

Nearly two-thirds of the lakes sampled in the WLS-I
are  located in federally designated  wilderness areas
to which motorized access is prohibited (Wilderness
Act of 1964).  Because  the  prescribed method  of
accessing lakes in the NLS was  by  helicopter
(Linthurst  et  al. 1986),  an alternative method  of
sampling wilderness area lakes was proposed by the
Forest Service to EPA: personnel from the Forest
Service  would serve  as ground  access sampling
teams and as  field site  managers.  Additionally,
because of concerns that data collected  by  ground
access methods would not be comparable to those
collected by helicopter access  methods,  the Forest
Service granted permission to sample 50 wilderness
lakes by both methods (45 of which  were  actually
sampled).  The results of this paired sampling were
used to determine whether or not the  data collected
by ground crews  needed to be calibrated against
those collected by helicopter crews.

The  WLS-I was designed to  provide information
to assess  the chemical status of  lakes in areas of
the  western United  States containing the majority
of lakes with low acid neutralizing capacity (Omernik
and  Griffith  1986).  Lakes were selected from five
subregions of the  West (see  inside  front  cover):
California  (Subregion 4A), Pacific  Northwest (Sub-
region 4B), Northern Rockies (Subregion 4C), Central
Rockies (Subregion  4D), and Southern  Rockies
(Subregion 4E). The boundaries of  these subregions
generally separated geomorphic provinces, and most
of the study areas were at high elevations. A number
of variables thought to influence  or be influenced
by surface water acidification were measured using
standardized methods.

-------
The WLS-I data base may be used to  investigate
correlative relationships among chemical variables
and to quantify the present chemical status of lakes
in areas of the  western United States  covered in
the survey. Additionally, the data  from  the  WLS-I
can be used with the ELS-I data to obtain a national-
scale assessment of the extent to which lakes in
the United States are potentially at risk  as a result
of present levels of acidic deposition.

The primary objectives of the WLS-I were to:

  1.  determine the percentage, by number and area,
     and  location  of lakes  that are acidic  in
     potentially  sensitive regions of the  western
     United States;

  2.  determine the percentage, by number and area,
     and  location  of  lakes that  have low  acid
     neutralizing capacity  in  potentially  sensitive
     regions of the western United States;  and

  3.  determine the  chemical characteristics of lake
     populations in potentially sensitive regions of
     the western United States and provide the data
     base for selecting lakes for future studies.

The  purpose  of this  report  is to  present results
obtained from the WLS-I. The data base generated
during the WLS-I will be used to frame subsequent
projects designed to examine short-term acidifica-
tion  as  a  result of  episodic events and long-term
acidification  through  an  evaluation  of trends  in
aquatic  ecosystems.

-------
                                          Section 2
                                           Methods
2.1  Design

2.1.1  Data Quality Objectives
The statistical design, sampling  and analytical
methods, and  quality assurance (QA) activities for
the Western Lake Survey -  Phase I (WLS-I) were
structured to meet specific data quality objectives
(DQOs;  Silverstein et  al.  1987a) for reporting
population estimates and chemical variability. The
DQOs were  used to  evaluate sampling, field
laboratory, and analytical laboratory performance.

The primary DQOs for the WLS-I (Table 2-1) were
defined  in  terms of  precision, accuracy,  and
detectability. Precision  was  expressed as percent
relative standard deviation. Accuracy was expressed
as maximum absolute bias, in percent. Detectability
was expressed as an expected value range and a
required detection limit. Each parameter measured
at the lake sampling  sites, in the field laboratories,
and in the analytical laboratories was evaluated with
respect to the  DQOs, for which appropriate  values
and ranges were established. The values and ranges
used for the WLS-I were based on experience gained
during the  Eastern Lake  Survey  -  Phase  I  (ELS-I;
Drouse et al. 1986) and on standard laboratory QA
requirements. Certain other DQOs were also
considered  in  the survey  design.  These included
completeness,  comparability, and  representative-
ness.

Completeness  was defined as a  measure  of the
quantity of acceptable data actually collected relative
to the total  quantity that was  attempted. For the
WLS-I, completeness was  targeted at or above 90
percent for all parameters. That is, of the probability
sample lakes selected (Section 2.2.1), 90 percent
were expected  to yield samples that would meet the
QA criteria  and  that could be used  in  defining
characteristics  of the  target population.

Comparability  of data is a measure by  which
similarity within  and  among  data sets can  be
established confidently.  Comparability can  be
considered in the context of comparisons among data
sets of different origin and also within components
of a single data set. Comparability of the WLS-I data
was ensured by requiring that  all  sampling  crews
and  laboratory analysts  use standardized  proce-
dures.  Comparability of the  WLS-I data  was
confirmed by checking the analytical results against
known standards and comparing the results to data
collected in previous  studies on  the same  lakes
(Section 7).

Representativeness,  another  measure of data
quality, is the degree to which sample data accurately
and  precisely reflect the characteristics of the
population of lakes. Representativeness of a subset
of sampled lakes can be established with complete
confidence only when the sample results can be
compared to the total population of lakes. Obviously,
representativeness of the WLS-I  data cannot be
established unequivocally because complete survey
data for all  lakes in the  West are unavailable.  In
the absence of this census  information,  the  best
assurance that the subset of lakes sampled repre-
sents the target population is obtained through the
implementation  of a sound,  statistically  based
sampling design (Section 2.2).

Additional confidence  in the representativeness  of
the sampled lakes is obtained by comparing results
from other statistically based surveys conducted  in
the same areas. The results of one such survey by
Bradford et al. (1968)  of  170 Sierra Nevada lakes,
and of a more recent survey of 124 of these same
170 lakes by Fox et al. (1982), are very similar  to
those obtained in the WLS-I (Sections  5 and 7). This
agreement  indicates, but does  not necessarily
confirm, the comparability and the representative-
ness of the WLS-I data.

2.7.2  Summary of the Statistical Design

During the WLS-I, one sample per lake was collected
during the  fall turnover  period from  the apparent
deepest part of the lake as an index to the essential
characteristics of each lake. The fall turnover period
was  selected because  lake water chemistry within
any  single  lake was expected  to  be the  most
homogeneous during this season.  Lakes were
selected using  a  stratified  design  with  equal
allocation  of sample  lakes to strata.  Lakes  were
selected in  each stratum by systematic sampling  of
an ordered  list following a random  start. The choice
of a desired sample size of 50 target lakes per stratum

-------
was based on the judgment that this sample size
would yield  adequate  precision  for  population
estimates by stratum.

2.7.3  Definition of Study Area

The population of lakes to be sampled was defined
as lakes located in those areas of the western United
States expected to contain  an  abundance of lakes
with alkalinity < 400 /ueq  L~1. The boundaries of the
western  region (Figure 2-1) were derived from  a
national  map of  surface  water  alkalinity (Omernik
and Powers 1983) and were considered to contain
from 95 to 99 percent of the lower alkalinity lakes
in the western United States. Unlike the subregions
sampled in the Northeast  and Upper Midwest in the
ELS-I (Linthurst et al.  1986), no subregions in the
West were conterminous, because the low alkalinity
areas of the West  generally coincide  with major
mountain ranges.
2.2  Lake Selection

2.2.1  Probability Sample
The stratification factors used in lake selection were
region, subregion, and alkalinity map class. The West
was defined as one region (Region 4 in the National
Lake Survey) with five subregions (4A through 4E).
Descriptive names  of  subregions were  assigned
(Figure 2-1). Lakes representing each of the three
alkalinity map  classes (ANC < 100, 100-199, 200-
400 fjeq L"1 derived from Omernik and Griffith 1986,
Figure 2-1) were  found within each  of the five
subregions; thus,  15 strata were defined  in  the
WLS-I.

Lakes were identified and listed using 1:100,000-
scale U.S.  Geological Survey  (USGS) topographic
maps in contrast to the ELS-I, in which 1:250,000-
scale maps were used  for lake selection.  Use of
1:100,000-scale maps  permitted lakes with  a
 Table 2-1.   Data Quality Objectives for Detectability, Precision, and Accuracy, Western Lake Survey - Phase I
Detectability
Site3
2, 3

3

3


3

3
1, 3

2


2, 3



3



3


3

3
3

Variableb
Al,
extractable

Al, total

Acid Neutral-
izing Capacity
(ANC)
Ca

cr
Conductance

True Color


Dissolved
Inorganic
Carbon (DIG)

Dissolved
Organic
Carbon (DOC)

F~, total
dissolved

Fe

K
Mg

Method0
Complexation with 8-
hydroxyquinoline and
extraction into MIBK
followed by AAS
(furnace)
AAS (furnace)

Titration and Gran
analysis

AAS (flame) or
ICPAES
Ion chromatography
Conductivity cell
and meter
Comparison to
platinum-cobalt
color standards
Instrumental (acid-
ification, C02
generation, IR
detection)
Instrumental (UV-
promoted oxidation,
CO2 generation, IR
detection)
Ion-selective
electrode and
meter
AAS (flame) or
ICPAES
AAS (flame)
AAS (flame) or
ICPAES
Laboratory
Reporting
Units
mg L"1

mg L"1

peq L~1


mg L~1

mg L"1
^S cm"1

PCUh


mg L~1



mg L-1



mg L"1


mg L"1

mg L"1
mg L"1

Expected
Ranged
0.005-1.0

0.005-1

-100 -


0.5 -

0.2 -
10 -

0 -


0.05 -



0.1 -



0.01 -


0.01 -

0.1 -
0.1 -


.0

1000


20

10
1000

200


15



50



0.20


5.0

1.0
7.0

Required
Detection
Limit (RDL)
0.005

0.005

f


0.01

0.01
g

0


0.05



0.1



0.005


0.01

0.01
0.01

Intralaboratory
Precision
%RSD
Upper Limit6
10(AI>0.01 mg L~1)
20IAK0.01 mg L~1)

10(AI>0.01 mg L~1)
20IAK0.01 mg L~1)
10


5

5
2

±5'


10



5(DOC>5 mg L'1)
10(DOC<5 mg L~1)


5


10

5
5

Accuracy
Maximum
Absolute
Bias(%)
10
20

10
20
10


10

10
5

i


10



10
10


10


10

10
10

                                                                               (continued)

-------
Table 2-1.   (continued)
Detectability
Site8
3

3
3

3
3


1, 2

3

3


3
2

Variableb
Mn

Na
NH4 +

N03-
P, Total


PH

pH

Si02


so4-2
Turbidity

Method0
AAS (flame) or
ICPAES
AAS (flame)
Automated color-
imetry (phenate)
Ion chromatography
Automated color-
imetry (phospho-
molybdate)
pH electrode and
meter
pH electrode and
meter
Automated color-
imetry (molybdate
blue)
Ion chromatography
Instrument
(nephelometer)
Laboratory
Reporting Expected
Units Ranged
mg L~1

mg L~1
mg L~1

mg L~1
mg L~1


pH units

pH units

mgL-1


mg L~1
NTUk

0.01 -

0.5 -
0.01 -

0.01 -
0.005 -


3 -

3 -

0.2 -


1.0 -
2 -

5.0

7.0
2.0

5
0.070


8

B

250


20.0
15

Required
Detection
Limit (RDL)
0.01

0.01
0.01

0.005
0.002


j

i

0.05


0.05
2

Intralaboratory
Precision
%RSD
Upper Limit6
10

5
5

10
10(P>0.01 mg L"1)
20(P<0.01 mg L"1)

±0.1'

±0.05'

5


5
10

Accuracy
Maximum
Absolute
Bias(%)
10

10
10

10
10
20

±0.10'

±0.10'

10


10
10

a 1 = lake site, 2 = field laboratory, 3 = analytical laboratory.
bDissolved ions and metals were determined, except where noted.
c Methods descriptions are given in Volume II (Tables 2-1 and  5-3), Hillman et al. (1986) and Kerfoot and Faber  (1987).
dRanges are for lake waters.
"Unless otherwise noted, this is the percent relative standard  deviation (% RSD) at concentrations greater than 10 times the required
 detection limit.
f Absolute blank value was required to be < 10 ^ieq L~1.
8The mean of six nonconsecutive blank measurements was required to be <0.9 /jS cm"1.
hAPHA platinum cobalt units.
' Absolute precision or accuracy goal in applicable units.
' Not applicable.
kNephelometric turbidity units.
minimum size of approximately 1 hectare (ha) to be
identified, as opposed to the approximately 4 ha used
in the ELS-I.

Strata boundaries were delineated, and all lakes on
the maps were numbered in spatial order within each
stratum. The final number of lakes identified in each
stratum was the total number of lakes in the map
population for the stratum. All population estimates
for physical and chemical attributes computed in this
study refer to the map population of lakes and  do
not necessarily represent conditions in lakes outside
the area of coverage  or in those  not depicted  on
the USGS topographic maps  used.  For example,
population estimates  cannot  be made  for  lakes
smaller than 1 ha.

Within each stratum, a systematic random sample
of 50 lakes was drawn. Lake numbers were entered
into a computer  file in numerical order as labeled
on the maps. In each stratum, the first lake was
selected at random between lakes 1  and k (where
k is the size of the map population divided by  the
desired sample size), and every kth lake was selected
thereafter. This sample is a true probability sample,
i.e., within each stratum, each lake had  an equal
probability of inclusion.

2.2.2  Identification of Non-target Lakes

"Non-target" lakes (Table 2-2) are those bodies of
water that either were not the focus of the survey's
objectives or could  not  be  sampled within  the
constraints of a  synoptic survey.  Non-target lakes
(Table  2-2) were first identified in the probability
sample by the examination of large-scale (1:24,000
or 1:62,500) USGS topographic maps. Categories of
non-target lakes identified by examining these maps
include:

 1.  No  lake present: lakes initially  identified on
     1:100,000-scale maps that did not appear on
     more detailed, larger scale maps.

 2.  Flowing  water:  sites  identified  as  lakes on
     1:100,000-scale maps that appeared as points
     on a stream on larger scale maps. However,
     if the small-scale maps were  more recent than

-------
Figure 2-1.   Subregions and alkalinity map classes ( • .• < 100 fjeq i"1; Q : 100-199 yeq L 1; D . 200-400 fjeg L~\ in study
            areas of the Western Lake Survey - Phase I.
                                    Northern
                                    Rockies (4C)
                                                                                 Alkalinity Map Classes
                                                                                       <100

                                                                                       100-199

                                                                                       200-400

                                                                                       Subregion
                                                                                       Boundary
                                                                   Central
                                                                   Rockies (4D)
      Pacific
      Northwest (4B)
    California (4A)
                                                                                          Southern
                                                                                          Rockies (4E)
     the large-scale maps and the lake in question
     was known to be a new reservoir, the lake was
     not eliminated.

 3.  Urban/Industrial/'Agricultural: lakes  sur-
     rounded  by or  adjacent  to intense  urban,
     industrial, or agricultural  land  use including
     tailing ponds, water treatment lagoons, andfish
     hatcheries.

 4.  Wetlands: lakes identified  on 1:100,000-scale
     maps that appeared as wetlands on larger scale
     maps.
 5.   Too small (<  1 ha):  lakes identified on
     1:100,000-scale  maps that were smaller than
     approximately 1 ha when measured on larger
     scale maps. Because the resolution of most
     1:100,000-scale  maps was  about 1 ha, this
     limit was established for consistency.

Following the elimination of non-target lakes by map
examination, it was necessary to restore the number
of selected lakes per stratum to 50. Also, in the event
that field crews encountered non-target  lakes,
additional lakes were selected.  These  additional

-------
lakes were chosen from the computer file of lakes
remaining  after  initial selection  using  the  same
procedure as for the original set of lakes, and were
also evaluated on large-scale maps to eliminate non-
target lakes.

The  lakes that met the selection  criteria after the
map  evaluation were provisionally designated as
"target" lakes. This designation was refined further
as a result of information obtained during or after
field sampling. The categories and numbers of non-
target lakes eliminated during or after sampling are
given in Table 2-2. The definitions of these categories
are as follows:
 1.  No lake present: sites visited that were found
    to be dry.

 2.  Flowing  water:  sites visited that were found
    to be streams.
3.
5.
6.
     High conductance: lakes visited that were found
     to have a measured conductance greater than
     1500/uScm"1.

     Urban/Industrial/Agricultural: lakes that were
     surrounded  by or were adjacent  to  intense
     anthropogenic activities.

     Too shallow:  lakes that were  too shallow
     (generally, less than 0.75 m) to obtain a clean
     (i.e., free of debris and sediment) sample.
     Other: lakes that were inaccessible due to a
     permanent feature of the lake that prevented
     helicopters from landing safely (e.g.,  power
     lines).
Lakes were classified as "not visited" (Table 2-2)
if the reason for not sampling was unrelated to a
permanent feature of the lake. This category of lakes
includes those that could not be  visited  because
permission to sample was not obtained or weather
 Table 2-2.   Non-Target Lakes and Lakes Not Visited or Not Sampled, Western Lake Survey - Phase I

 A. Probability sample lakes determined to be non-target from large-scale map examination
California
Categories
No Lake Present
Flowing Water
Urban/lndustria I/Agricultural
Wetlands
Too Small « 1 ha)
4A
7
1
2
2
14
Pacific
Northwest
4B
0
0
1
4
24
Northern
Rockies
4C
1
0
0
2
7
Central
Rockies
4D
1
1
0
1
18
Southern
Rockies
4E
1
1
3
0
3
Total
10
3
6
9
66
 Total
                           26
                                        29
                                                      10
                                                                   21
                                                                                             94
 B. Non-target probability sample lakes determined during or after sampling
California
Categories
No Lake Present
Flowing Water
High Conductance
Urban/lndustnal/Agncultural
Too Shallow
Other
4A
7
0
0
1
10
0
Pacific
Northwest
4B
2
0
0
0
15
0
Northern
Rockies
4C
5
0
1
0
18
0
Central
Rockies
4D
6
0
0
0
14
2
Southern
Rockies
4E
2
0
1
0
13
1
Total
22
0
2
1
70
3
 Total
                           18
                                        17
                                                     24
                                                                   22
                                                                                17
                                         98
 C. Probability sample lakes that were not visited
Categories
No Access Permit
Bad Weather
Wrong Lake
Frozen
California
4A
2
0
0
1
Pacific
Northwest
4B
4
1
0
0
Northern
Rockies
4C
0
0
1
7
Central
Rockies
4D
0
1
0
21
Southern
Rockies
4E
2
0
0
20
Total
8
2
1
49
Total
              22
                                                                                22
                                                                                             60

-------
conditions prohibited access. Lakes that were visited
but were not sampled because they were frozen or
were sampled, but were found during data validation
(Section  3.3) to  be the  wrong lake,  were also
classified as not visited. The target or non-target
status of this group of lakes could not be determined;
thus, they represent incompleteness in the sample.
For  statistical analyses,  it was  assumed that
lakes not visited were a random subsample  of the
original probability sample  and thus had the same
proportion of non-target lakes as the lakes that were
visited. This assumption appears valid because no
systematic differences in lake area,  watershed area,
or elevation were found between frozen lakes  and
sampled lakes (Section 5.2.2).

However, some differences were noted in the Central
and Southern Rockies  between the frozen  lakes and
the  sampled lakes (Table  2-3).  Frozen  lakes were
generally smaller  and were  located at  higher
elevations than were  the sampled  lakes.  It is also
probable that the frozen lakes were shallower than
the  sampled lakes,  although no measurements of
lake depth in the frozen lakes are available to confirm
this assumption. Analyses shown later in this report
(Section 6.4.2) indicate that ANC is inversely related
to elevation in the Central Rockies. That frozen lakes
in this subregion were generally located at elevations
higher than those estimated from the sampled lakes
for the target population suggests that the popu-
lation estimates for the Central Rockies are not based
on a random subsample. The degree to which the
population estimates underestimate the number of
low ANC lakes in the  Central Rockies is difficult to
assess. However, comparison of WLS-I results with
previous surveys shows relatively close agreement
(Section 7.4).

The population estimates for the number of low ANC
lakes in the  Southern  Rockies are probably not
affected to the degree that may have occurred in
the  Central Rockies. The  differences in  lake  size
between frozen and sampled lakes is smaller and
there is no apparent relationship between lake ANC
and lake elevation in Subregion 4E (Section  6.4.2).
               However, with such a large proportion of the selected
               subsample unavailable for sampling in the Central
               and Southern Rockies, it is not possible to determine
               confidently the impact that frozen lakes had on the
               characterizations of the lake populations.

               2.2.3  Identification of Wilderness Lakes

               All  lakes  in Forest  Service  wilderness  areas  and
               some lakes in national parks could be sampled only
               by ground crews  because of legislation restricting
               the use of motorized vehicles (Wilderness Act of
               1964). To  make plans for sampling access, all of the
               selected lakes that were not eliminated as non-target
               by map examination were evaluated to determine
               if they were located within the boundaries of a Forest
               Service wilderness  area  or a national park. These
               evaluations were based on consultations with Forest
               Service and National Park Service personnel and on
               close  examination  of the  most  accurate  maps
               available.  The wilderness lake classifications may
               not be entirely correct,  however, particularly  for
               newly created wilderness  areas for which  the
               congressionally  certified  boundaries  were  not
               available at the time wilderness lake identifications
               were  made.  The names and  locations of lakes
               sampled in Forest  Service  wilderness  areas  and
               national parks are given in Figures 2-2 and 2-3. The
               number of lakes sampled in  each area is given in
               Table 2-4.

               2.2.4  Special Interest Lakes

               Other lakes, in addition to  those chosen in  the
               probability sample, were included in the  WLS-I.
               Forty-two lakes that were not selected randomly and
               are  or were  the subjects  of  relevant research
               programs were selected as special interest lakes.
               Of these lakes, samples were collected from 32. All
               western  lakes  in  the  current EPA Long-Term
               Monitoring Program  were selected as special
               interest  lakes.  Other special interest lakes were
               included based on recommendations from state and
               federal agencies. Data from  these lakes were  not
               used in  computing population estimates, but  the
 Table 2-3.   Population Estimates of Median Lake Area and Median Elevation for Frozen Lakes and Sampled Lakes in Central
           Rockies (4D) and Southern Rockies (4E), Western Lake Survey - Phase I
 Stratum
                       Target Lakes Visited (n**»)
                  Lake Area (ha)
                        frozen
sampled
                                                  frozen
sampled
                  Elevation (m)
                                                                           frozen
sampled
Central Rockies
4D1
4D2
4D3
Southern Rockies
4E1
4E2
4E3

11
7
3

14
5
1

43
47
39

46
52
41

3.5
2.4
6.1

2.6
2.6
1.7

5.6
5.5
3.5

3.0
3.3
3.5

3338
3042
2865

3394
3237
3188

3219
3042
2687

3307
3456
3147

-------
Figure 2-2.     Locations of Forest Service wilderness areas ( S ) and national parks ( ® ) containing lakes sampled in  the
               California (4A) and Pacific Northwest (4B) subregions, in study areas of the Western Lake Survey - Phase I.
                                20 22  19
           28
4A
         I  Wilderness Areas
         | National Parks

         —«^_*
           Subregion Boundary
                                                                       Pacific Northwest (4B)

                                                                 19. Pasayten Wilderness
                                                                 20. Mount Baker Wilderness
                                                                 21. Noisy Diobsud Wilderness
                                                                 22. North Cascades National Park
                                                                 23. Lake Chelan-Sawtooth Wilderness
                                                                 24. Boulder River Wilderness
                                                                 25. Glacier Peak Wilderness
                                                                 26. Henry M. Jackson Wilderness
                                                                 27. Buckhorn Wilderness
                                                                 28. Olympic National Park
                                                                 29. Alpine Lakes Wilderness
                                                                 30. William 0. Douglas Wilderness
                                                                 31. Clearwater Wilderness
                                                                 32. Mount Rainier National Park
                                                                 33. Goat Rocks Wilderness
                                                                 34. Indian Heaven Wilderness
                                                                 35. Columbia Wilderness
                                                                 36. Mount Hood Wilderness
                                                                 37. Mount Jefferson Wilderness
                                                                 38. Mount Washington Wilderness
                                                                 39. Three Sisters Wilderness
                                                                 40. Waldo Lake Wilderness
                                                                 41. Diamond Peak Wilderness
                                                                 42. Sky Lakes Wilderness
                                                                                California (4AJ

                                                                           1.  Siskiyou Wilderness
                                                                           2.  Marble Mountain Wilderness
                                                                           3.  Trinity Alps Wilderness
                                                                           4.  Thousand Lakes Wilderness
                                                                           5.  Lassen Volcanic National Park
                                                                           6.  Caribou Wilderness
                                                                           7.  Bucks Lake Wilderness
                                                                           8,  Desolation Wilderness
                                                                           9.  Mokelumne Wilderness
                                                                          10.  Hoover Wilderness
                                                                          11.  Emigrant Wilderness
                                                                          12.  Yosemite National Park
                                                                          13.  Minarets Wilderness
                                                                          14.  John Muir Wilderness
                                                                          15.  Kaiser Wilderness
                                                                          16.  Dinkey Lakes Wilderness
                                                                          17.  Kings Canyon National Park
                                                                          18.  Sequoia National Park

-------
Figure 2-3.     Locations  of Forest Service wilderness areas ( Q ) and national parks ( @ )  containing lakes sampled in the
               Rocky Mountain subregions: the Northern Rockies (4CJ. the Central Rockies (4D), and the Southern Rockies
               (4E). in study areas of the Western Lake Survey - Phase I. Lakes located in roadless areas are also shown (\3).

                                                                                   Northern Rockies I4C)
                                                                                    43.
                                                                                    44.
                                                                                    45.
                                                                                    46.
                                                                                    47.
                                                                                    48.
                                                                               Glacier National Park
                                                                               Great Bear Wilderness
                                                                               Cabinet Mountains Wilderness
                                                                               Mission Mountains Wilderness
                                                                               Rattlesnake Wilderness
                                                                               Anaconda-Pintlar Wilderness
                                                                           49. Selway-Bitterroot Wilderness
                                                                           50. Gospel Hump Wilderness
                                                                           51. River of No Return Wilderness
                                                                           52. Hell's Canyon Wilderness
                                                                           53. Eagle Cap Wilderness
                                                                           54. Sawtooth Wilderness
                                                                                              I
        Central Rockies (4D)

        55. Lee Metcalf Wilderness
        56. Absaroka-Beartooth Wilderness
        57. Cloud Peak Wilderness
            Yellowstone National Park
            Teton Wilderness
            Grand Teton National Park
        61. Jedidiah Smith Wilderness
        62. Gros Ventre Wilderness
        63. Fitzpatrick Wilderness
        64. Wind River Roadless Area
        65. Popo Agie Wilderness
        66. Bridger  Wilderness
        67. High Uintas Wilderness
        68. Lone Peak Wilderness
58.
59.
60.
                   \
                      \
                         \
                            \
                                   Southern Rockies (4E)

                                   69. Mount Zirkel Wilderness
                                   70. Comanche Peak Wilderness
                                   71. Never Summer Wilderness
                                   72. Rocky Mountain National Park
                                   73. Indian Peaks Wilderness      	!""
                                   74. Flat Tops Wilderness             \
                                   75. Eagles Nest Wilderness          '
                                   76. Mount Evans Wilderness         I
                                   77. Holy Cross Wilderness           \
                                   78. Hunter Frying Pan Wilderness
                                   79. Collegiate Peaks Wilderness
                                   80. Maroon Bells-Snowmass Wilderness
                                   81. West Elk Wilderness
                                   82. Weminuche Wilderness
                                   83. South San Juan Wilderness
                                   84. Wheeler Peak Wilderness
                                                                                                 Wilderness Areas

                                                                                                 National Parks
                                                                                                 Roadless Area

                                                                                                 Subregion Boundaries
                            10

-------
Table 2-4. Forest Service Wilderness Areas and National Parks Containing Probability Sample Lakes, Grouped by
Subpopulations Used in Data Analysis, Western Lake Survey - Phase 1



1.
2.
3.


4.
5.
6.
7.


8.
9.
10.
11.
12.
1 ^
I O.

14.
15.
16.
17.
18.
Total



19.
20.
21.
22.
23.
24.
25.
26.
27.
28.

29.
30.
31.
32.
33.
34.

35.
36.
37.
38.
39.
40.
41.
42.
Total

Subpopulation
California (4A)
Klamath Mountains (CA)
Siskiyou Wilderness
Marble Mountain Wilderness
Trinity Alps Wilderness

Southern Cascades (CA)
Thousand Lakes Wilderness
Lassen Volcanic National Park
Caribou Wilderness
Bucks Lake Wilderness

Tahoe Area (CA)
Desolation Wilderness
Mokelumne Wilderness
Yosemite Area (CA)
Hoover Wilderness
Emigrant Wilderness
Yosemite National Park
IV/T nm+ lA/iM norte?
Minarets vvnuerness
Southern Sierra Nevada (CA)
John Muir Wilderness
Kaiser Wilderness
Dinkey Lakes Wilderness
Kings Canyon National Park
Sequoia National Park
(4A)

Pacific Northwest (4B)
North Cascades Area (WA)
Pasayten Wilderness
Mount Baker Wilderness
Noisy Diobsud Wilderness
North Cascades National Park
Lake Chelan-Sawtooth Wilderness
Boulder River Wilderness
Glacier Peak Wilderness
Henry M. Jackson Wilderness
Olympic Mountains (WA)
Buckhorn Wilderness
Olympic National Park
Alpine Lakes (WA)
Alpine Lakes Wilderness
Washington Cascades (WA)
William O. Douglas Wilderness
Clearwater Wilderness
Mount Rainier National Park
Goat Rocks Wilderness
Indian Heaven Wilderness
Oregon Cascades (OR)
Columbia Wilderness
Mount Hood Wilderness
Mount Jefferson Wilderness
Mount Washington Wilderness
Three Sisters Wilderness
Waldo Lake Wilderness
Diamond Peak Wilderness
Sky Lakes Wilderness
(4B)

Lakes
Sampled

13
2
4
7

13
2
7
3
1

14
11
3
19
1
4
9
5

38
16
1
1
11
9
97


28
7
1
1
6
2
1
8
2
4
1
3
24
24
13
5
1
5
1
1
21
1
1
2
1
8
2
1
5
90




43.
44.
45.
46.
47.


48.
49.
50.
51.


52.
53.
54.

Total




55.
56.

57.

58.
59.
60.
61.
62.


63.
64.
65.
66.

67.
68.
Total



69.

70.
71.
72.
73.


74.


75.
Subpopulation
Northern Rockies (4C)
Northwest Montana (MT)
Glacier National Park
Great Bear Wilderness
Cabinet Mountains Wilderness
Mission Mountains Wilderness
Rattlesnake Wilderness

Bitterroot Area (MT/ID)
Anaconda-Pintlar Wilderness
Selway-Bitterroot Wilderness
Gospel Hump Wilderness
River of No Return Wilderness

Blue Mountains (ID/OR)
Hell's Canyon Wilderness
Eagle Cap Wilderness
Sawtooth Mountains (ID)
Sawtooth Wilderness

(4C)


Centra/ Rockies (4D)
Southern Montana (MT)
Lee Metcalf Wilderness
Absaroka-Beartooth Wilderness
Cloud Peak (WY)
Cloud Peak Wilderness
Northwest Wyoming (WY)
Yellowstone National Park
Teton Wilderness
Grand Teton National Park
Jedidiah Smith Wilderness
Gros Ventre Wilderness

'Wind River Mountains (WY)
Fitzpatrick Wilderness
Wind River Roadless Area3
Popo Agie Wilderness
Bridger Wilderness
Wasatch/Uinta Mountains (UT)
High Uintas Wilderness
Lone Peak Wilderness
(4D)

Southern Rockies (4E)
Mount Zirkel (CO)
Mount Zirkel Wilderness
Rocky Mountain Park Area (CO)
Comanche Peak Wilderness
Never Summer Wilderness
Rocky Mountain National Park
Indian Peaks Wilderness

Flat Tops (CO)
Flat Tops Wilderness

Central Colorado (CO)
Eagles Nest Wilderness
Lakes
Sampled

11
5
1
1
2
2

44
2
31
1
10

10
2
8
17
17

82



16
1
15
9
9
11
6
1
2
1
1

44
5
6
7
26
16
15
1
96


17
17
36
1
1
22
12

8
8
1 1
\lt
2

-------
Table 2-4.   (continued)
    Subpopulation
 Lakes
Sampled
76. Mount Evans Wilderness
77. Holy Cross Wilderness
78. Hunter Frying Pan Wilderness
79. Collegiate Peaks Wilderness
80. Maroon Bells-Snowmass
Wilderness
81. West Elk Wilderness
S. Colorado/New Mexico
(CO.NM)
82. Weminuche Wilderness
83. South San Juan Wilderness
84. Wheeler Peak Wilderness
Total (4E)
2
2
2
1
2

1
17

14
2
1
90

Total in the West
442
'Located on an  Indian Reservation and technically not a
 wilderness area, but included here to complete the Wind River
 Range (Figure 2-3).

results are summarized in Table 5-23 and individual
lake results are shown in  Volume  II (Eilers et al.
1987b).

2.2.5   Final Lake Lists and Maps

Lake names, identification (ID) numbers, geographi-
cal coordinates, wilderness  area or park names, and
map names were entered  into computer files and
were printed for field crews. If no name was printed
on the  map for a given lake, the entry in the file
was "no name." Each lake was assigned a unique
ID number coded for the stratum in which it occurred
(e.g., 4A2-011  is the 11th lake selected in alkalinity
map class 2 of Subregion A in  Region 4). The latitude
and longitude for each lake were measured with 11 -
point dividers  to the  nearest degree, minute, and
second and were checked by visual examination and
computer generated overlays. Additionally, ID codes
were printed on the  topographic maps for use  by
field crews in locating the lakes.

2.3   Design Applications and
 Restrictions

2.3.1   Extrapolation from Sample to Population
For statistical  analyses, the probability  sample was
treated as a simple  random sample within each
stratum. The ordering  of the lakes and the systematic
selection process were designed to increase the
precision over that which would have been obtained
with a simple random  sample; therefore, statements
of precision for population  estimates are probably
conservative.
When population estimates from combined strata are
required (e.g., when making subregional or regional
estimates), expansion factors or weights (W) must
be  used  because the sampling  intensity varied
among strata. These  weights (the stratum target
population size divided by the stratum sample size)
vary considerably among strata; for example, lakes
sampled in one stratum may each represent three
lakes, whereas each lake sampled in another stratum
may represent 36 lakes.
This design permits strata to be combined when it
is meaningful to do so. The flexibility of the design
also allows estimates to be  made for  specific
subpopulations, or subsets of lakes; however, unless
the definition of the population of.interest is clearly
stated, conclusions based on the WLS-I data  can
be misleading.

2.3.2  Estimating the Target Population Size and
Attributes

The first steps in the statistical analysis of the data
were to estimate the target population size in each
stratum ( N; Figure 2-4) and to determine the stratum-
specific weights needed for interstratum estimation.
The target population size  in each  stratum  was
estimated in two units: number of lakes and lake
area (in hectares).
The statistical frame is defined by the list of lakes
identified by the map population. The map population
consists of both  target and non-target populations.
Within each stratum, it is possible to estimate the
size of the target  population by multiplying the
number of sampled lakes classified as target (n***)
by  the stratum-specific weight (W). This  weight is
the inverse of the inclusion probability (P) of a target
lake  in the  final sample,  which  is determined
according to the following equation:

               P=  1/W = (n*/N*)q            (1)

  where n* = the size of the sample  drawn from
              the map population

         N* = the size of the map population

          q = the probability that a target lake in
              the drawn sample is actually visited;
              computed by dividing the actual
              number of lakes visited (n* - nnb - n0)
              by the number of lakes intended to
              be visited (n* - nnb)

              where nnb = the number of non-
                          target lakes in the origi-
                           nal sample, as deter-
                           mined from the maps,
                          and
                      n0  =  the number of lakes not
                           visited.
                      12

-------
Figure 2-4.   Procedures used to estimate the target population size. Western Lake Survey - Phase I.
                                                                             Map Population
                                                                                 IN")
                    Large Scale Map Examination
                                                                             Selected Lakes
                                                                                 (n*)
                         Estimated
                         Non-Target
                         Lakes in
                         Not Visited
                         Lakes
                                                                  I
                                                             Lakes Scheduled for
                                                             Visitation (n* -n^)
         Estimated
         Non-Target
         Lakes
         Selected
        I
                                                 Direct
                                               Examination
                                                                                  I
                                                                             Lakes Visited
                                                                             (n* -ftnb- na)
             N*
Multiply by     —
             /?*
                                                                  1
                                                             Target Lakes
                                                             Visited 
-------
                  A  =W(IA)
The variances of N and A for single  strata, were
estimated by:

V(N) = N* [(N* - n')/(n'-1)] [n** Vn'] [(n' - n***)/n']   (4)
(3)    where summation is over the appropriate subset of
      sample lakes in the appropriate strata, and where
      the values of Ware assigned according to the stratum
      in which the lake belongs.
V(A) = N*[(N* - n')/(n'-1)] [1 /n'] [IAZ - (ZA)Vn']   (5)
where n', the "effective sample size," is used in place
of n* because of incomplete visitation (i.e., n' = qn*).
The standard errors  are calculated as  the square
roots of the variances.

For estimates of populations covering multiple strata,
estimates and variances must  be computed within
strata and added or else computed with equations
containing weights (see  below).  Any  explicitly
defined subset of the total population of target lakes
in the West is a subpopulation. Subpopulations can
be defined over  any combination of  strata.  For
example, for any given variable, each observed value
of a  variable, X, defines  a subpopulation of lakes
having a value x  less than or  equal to that value.
Subpopulation  definitions  also  can be based on
geographic boundaries such as  states or national
parks. This procedure was used  in identifying specific
subpopulations in geomorphic units (Section 7).

Estimates for subpopulations that are defined within
single strata can be generated using formulae  that
are modifications  of the  equations given  above
(mathematically identical  to the algorithms used in
generating all  the statistics  for  the  survey). To
generate  single stratum  subpopulation equations
from equations 2 through 5, each n*** is replaced
by nz and each I by Iz, where nz is the number
of sample lakes in the subpopulation z and Zz is the
summation over the sample lakes in the subpop-
ulation z.

For example,  all  22  lakes (nz)  sampled in  Rocky
Mountain National Park are in Stratum 4E1, which
has W = 3.261  (Section 5.2.3,  Table 5;5). Thus, the
estimated target  population  size (N) for  Rocky
Mountain National Park  is (3.261)(22) = 72 lakes.
The 22 lakes have a combined area of 106.8 ha (IZA);
thus, the estimated area (A) of  the target population
in Rocky  Mountain National Park is (3.261)(106.8
ha) = 348.3 ha.

 A useful generalization, appropriate for any subpop-
 ulation and any combination of strata, is that.
      The use of equation  6 can be illustrated with the
      data  for population estimates  of acid neutralizing
      capacity (ANC) by state (Section 5.4.1.3, Table 5-10).
      In Utah, 30 lakes were sampled. Eleven of these
      were in Stratum 4D1 with W= 18.356, 13 in Stratum
      4D2 with W = 19.744, and 6 in Stratum 4D3  with
      W = 14.918 (Section 5.2.3, Table 5-5). The estimated
      target population size (N) for Utah  is calculated by
      adding the product of nz and  W  for each stratum,
      which is  the same as 11(18.356)  + 13(19.744) +
      6(14.918) = 548 lakes. A subpopulation of the lakes
      in Utah can be defined by a particular value for ANC.
      Only  one  of the lakes sampled in  Utah had an  ANC
      value  < 50 /ueq L~1. This lake was  in Stratum 4D2
      with a weight of 19.744; thus, the  estimated number
      of lakes in Utah with ANC < 50 fjeq L"1 is 20, and
      the estimated proportion is 20/548  = 0.04.

      Afurther generalization, used in data analysis,  leads
      to  a  similar formula  for the estimated variance of
      any  variable, X,  over any subpopulation  and
      combination of strata:
            Variance(X) = IWX2/IW - (IWX/IW)2    (7)
             N = IW, and A= IWA,

                       14
      where the set of sample lakes in the summation
      defines the subpopulation of  lakes for  which the
      variance is estimated.

      The  weighting factors are extremely  important.
      Estimating population parameters from sample data
      without accounting for weights can lead to erroneous
      calculations and incorrect interpretation. Examining
      relationships among variables  with the expectation
      that  these relationships  are representative of the
      population should only be done within strata  or by
      using weighting factors.

      By a method equivalent to calculating subpopulation
      estimates (number or area of lakes with concentra-
      tion less than or equal to x) and their associated
      upper confidence limits for all possible values of X,
      cumulative frequency  distributions [F(x>] and
      cumulative areal distributions [G(x>] were calculated.
      At any value x, these curves represent the estimated
      number or area of lakes in the population having
      a  value for that variable less than or equal to x, with
      the  95  percent  upper  confidence  limit  for that
      number (Section 5.1). For some variables, interest
      is in the number of lakes with concentrations above
      a  particular value (e.g., sulfate  > 50 (jteq L'\ so the
(6)    inverses of the cumulative frequency  distributions

-------
[1-F(x)J and  of the  cumulative  areal distributions
[1 -G(x)] were generated in a similar manner.

Quintiles and medians  for these cumulative fre-
quency and areal distributions were also calculated.
The quintiles (Qi through Q4) reflect the estimated
values of x separating the distribution into five equal
parts  (i.e., the 4th quintile is the  80th  percentile).
The median is the estimated value of x such that
half the lakes in  the population are  characterized
by concentrations of the variable  equal to or less
than the value of x.

2.3.3  Restrictions

The use  and  interpretation  of any  data set are
restricted  by the design,  the  quality of the data
obtained, and the sampling protocols. The map and
target populations and the period of  sampling are
the primary design  considerations influencing the
proper interpretation of the WLS-I data.

Estimates of the number of lakes within an area are
strongly affected  by the map scale used to define
the map population. Use of larger  scale  maps
provides greater resolution and allows smaller lakes
to be identified and included in the  population  under
consideration. The map scale  used to define the map
population  in  the WLS-I  was 1:100,000,  which
identified lakes as small as approximately 1 ha. Lakes
identified from 1:24,000-scale maps can be as small
as 0.1 ha. Some estimates of the  number of lakes
in portions  of the West have been based on  larger
scale  maps such as  1:62,500 or 1:25,000 (Turk and
Adams 1983).


To illustrate the influence of map scale in estimating
the total  number  of lakes, the number of lakes on
1:100,000-scale  maps within several areas  of
interest  were counted. The  process was then
repeated using larger scale maps covering the same
area as the 1:100,000-scale maps. The results  (Table
2-5) show that large-scale maps display more  lakes.

Comparisons of the results of the WLS-I to  those
from other lake surveys in the West must be done
with knowledge of the map scales used to prepare
the population estimates. No direct conclusions can
be drawn about the population of lakes less than
1 ha from WLS-I  data. This restriction also applies
to other categories of non-target  lakes defined  in
the WLS-I.

The period  of sampling restricts the conclusions of
the WLS-I  to the fall of 1985.  The  accuracy  of
extrapolating the  fall index sample (Section  2.1.2)
to other times of the year or to other years  is not
known.
2.4  Lake Characterization

Because of their effects on surface water chemistry,
several physical lake and watershed parameters are
important to consider when interpreting  chemical
data from lakes. These parameters were measured
or determined from  maps as described  in  the
following sections.

2.4.1  Lake Area

Lake area (in hectares) was measured  in triplicate
with an electronic planimeter  on  1:24,000-  or
1:62,500-scale topographic  maps. Large lakes or
reservoirs which occurred on more than one large-
scale map were measured on 1:100,000-scale maps.
Islands and wetlands within the lake were included
in the  measurement  of  watershed area (Section
2.4.4) rather than lake area.

The  following criteria were used  in  delineating
boundaries when a lake was  connected to other
bodies of water, was divided between adjoining
maps, or did not appear the same at different scales
of resolution:


 1.  the  boundary of  a lake was  defined to be at
     the narrowest point of its inlet or outlet;

 2.  if part of a lake was not shown on an adjoining
     map,  the perimeter  was  estimated  using
     topographic contour lines; and

  3.  when lakes appeared  as two distinct water
     bodies on a 1:100,000-scale map, but as one
     on  larger scale  maps, the entire water body
     was measured as if  it were one lake.

2.4.2  Elevation

The elevation of each lake was recorded in feet from
the  largest  scale map available  and was  subse-
quently converted to meters. When no lake elevation
was marked  on  the map, it was  estimated  by
interpolation of topographic contour lines.

2.4.3  Hydrologic Lake  Type

 Each lake was classified  by hydrologic type  (Wetzel
 1983) through visual examination of its morphome-
 try on the largest scale topographic maps available.
 "Seepage" lakes were defined as those having no
 inlet or outlet. "Closed" lakes were those with inlets
 but  no  outlets. "Drainage"  lakes were those with
 outlets, regardless of the presence or absence of
 inlets. "Reservoirs" were artificial lakes, as indicated
 by a dam at  an  outlet.  The reservoir class could
 include dammed river valleys and natural lakes with
 artificial outlet control structures.
                                                                         15

-------
Table 2-5.   Number of Lakes  Shown on Small-scale (1:100,000) and Large-scale (1:62,500; 1:24,000) Maps.
          Western Lake Survey - Phase I
Subregion
Northern Rockies
Central Rockies

Central Rockies

(4C)
(4D)

(4D)
Mountain
Range
Beartooth
Wind River

Uinta
1:100,000
Map Name
Alpine Lakes
Mt. Bonneville
Fremont Lake, N
Kings Peak
Number of Lakes
1:100,000
125
50
26
41
1:62,500
350
—
-
-
1:24,000
-
200
160
230
Ratio of
Large Scale to
Small Scale
2.8
4.0
6.1
5.6
2.4.4  Watershed Area

An electronic planimeter  was  used to measure
watershed area.  Watershed areas greater than 10
to 15  mi2 (approximately  26 to 39 km2) were
measured on 1 -.100,000-scale topographic maps, if
available;  otherwise, 1:250,000-scale topographic
maps were  used. Smaller watersheds were mea-
sured on larger scale maps.

2.4.5  Land Use/Land Cover

The percent of watershed area in each  of 11 land
use categories was determined from land use maps.
The following categories (Anderson et al. 1 976) were
used: urban  or  built-up  land,  agricultural land,
rangeland, deciduous forest land, evergreen forest
land,  mixed  forest  land,  surface  water, wetland,
barren land, tundra, and perennial snow or ice. These
categories are  represented  on  1:250,000-scale
USGS Land Use and  Land Cover overlays,  which
were available for all areas except parts of the Rocky
Mountains (most of the Northern Rockies and four
lakes in the Central Rockies). In these areas, land
use  information for  publicly owned  land was
obtained  from the  Forest Service,  National Park
Service,  Soil Conservation  Service,  and state
agencies.  Watersheds on private  land were char-
acterized  by interpreting  1:250,000-scale  USGS
topographic  maps and potential natural vegetation
maps (U.S.  Soil  Conservation Service 1976; U.S.
Geological Survey 1970).

Watersheds  were  delineated on  1:250,000-scale
USGS topographic maps and then were traced onto
the  land  use maps.  For most watersheds,  an
equidistant dot grid appropriate to the size  of the
watershed was placed over the land use map. The
sample lake was shaded so its area could be excluded
from the land use estimate. All the dots in each land
use  class were  counted; then the  template was
moved and  the  dots  were  recounted. The total
number of dots for each class was then divided by
the total number of  dots in the watershed to obtain
the percentage of each class in  the  watershed. If
the  watershed  contained  few  categories,  the
percentage of each was estimated visually.
2.4.5  Slope

Slope was calculated as the median of percent slope
(vertical distance divided by horizontal distance times
100) for approximately ten equidistant points within
each  lake's watershed. Slope measurements  for
watersheds greater than about 10-15 mi2 (26-39
km2) were taken from 1:100,000- or 1:250,000-scale
USGS topographic maps; measurements for smaller
watersheds were taken from larger scale maps.

An equidistant dot grid appropriate to the map scale
was placed over  the  watershed.  Percent slope
readings  were taken from a slope calculator placed
perpendicularly to the contour lines at the points.
For watersheds divided between two or more maps,
the number of calculations per map corresponded
to the proportion of the watershed on that map.

2.5  Sampling  and Analytical Methods

Sampling  for the  WLS-I was conducted from  10
September to  4 November  1985.  Except  where
permission to land a helicopter was denied, lakes
were accessed by helicopters with fixed floats and
were sampled by EPA and Lockheed Engineering and
Management Services Company,  Inc. (Lockheed-
EMSCO)  personnel. Most  lakes in  Forest  Service
Wilderness Areas, some lakes in National Parks, and
one other lake were accessed by ground  and were
sampled from inflatable boats by Forest Service and
Lockheed-EMSCO personnel. Water samples were
delivered for  initial  analyses, processing,  and
preservation to mobile  field laboratories located at
five field stations: Missoula and Bozeman, Montana;
Aspen, Colorado; Carson City, Nevada; and  Wenat-
chee, Washington.  Where  necessary, remote base
sites were established to collect samples from lakes
located at distances greater than approximately 150
miles (250 km) from  a field station. Regardless of
whether the helicopter crews were  operating from
the field station or from a remote base site, helicopter
crew  samples  were delivered  for field laboratory
analyses  within 16 hours of collection.  Sixty-two
percent of the samples collected  in the Survey by
ground crews  were also delivered  to  the field
laboratory on the same day they were sampled and

-------
were processed within 16 hours. Ninety percent of
all the  ground crew samples were delivered to the
field laboratory no later than on the day following
sample collection. After processing, samples were
shipped by overnight courier to analytical laborato-
ries for completion of analysis.

2.5.1  Field Sampling Activities

Field sampling activities conducted by ground crews
and helicopter crews included  site description and
location, as  well  as collection  of in situ lake data
and water  samples (Figure  2-5). Detailed field
sampling  protocols are  provided in  Bonoff and
Groeger (1987). A comparison of data collected using
both protocols is provided in Section 4.5.

2.5.1.1  Site Description and Location—Because of
its capability for prolonged high-altitude flying,  the
Aerospeciale Lama 315B  helicopter was  used to
sample the WLS-I lakes at all but one base site. In
Wenatchee,  Washington, a Bell 206 Long Ranger
was used as in the ELS-I (Morris et al. 1986). Lakes
to be sampled from helicopters were identified from
the air using  a  Loran-C  guidance  system and
topographic  maps. Aerial photographs  of the lakes
were  taken  and watershed  disturbances  were
recorded on  a standardized form. If  a lake was
accessible from the air, the helicopter was landed
at what appeared  to be the deepest part of the lake.
The deepest  point near the landing area was located
with an electronic depth finder.

Ground sampling personnel used topographic maps
and their familiarity with the local area to locate lakes
(Peck et al. 1985). Site locations were verified using
a number of methods, including comparing observed
lake morphometry  to that  shown  on  the map,
identifying  topographic  features, determining
position by  compass  triangulation,  and  enlisting
assistance from local experts. The deepest point of
the lake was estimated using shoreline topography
and soundings taken with the calibrated anchor line.

2.5.1.2  In  Situ  Measurements and Sample Col-
lection—Standardized field data forms were used by
helicopter crews and by ground crews to record  the
values of in situ measurements. Helicopter sampling
crews  used the fixed floats as platforms from which
to take in situ measurements and to collect water
samples.  Field  measurements taken  from the
helicopter were site depth, Secchi disk transparency,
temperature, pH, and conductance. Site depth was
determined with a weighted sounding line. Water
transparency was  measured with a 20-cm-diameter,
black-and-white Secchi disk. Helicopter crews used
Hydrolab  4041  units retrofitted with  glass  pH
electrodes and Beckman Lazarin reference elec-
trodes  to determine in situ  values of temperature,
pH, and conductance. The  Hydrolab  units  were
calibrated daily  with  a solution of 0.001  N KCI
(specific conductance = 147 /jS  cm 1 at 25°C) and
National  Bureau of  Standards (NBS) traceable pH
buffer solutions (pH = 7.00 and 4.01). Each day before
and after sampling,  temperature, pH, and conduc-
tance functions were also checked to ensure proper
instrument operation.  Temperature  was checked
against an NBS-traceable  thermometer. Conduct-
ance and pH functions were checked using  a COz-
saturated solution (pH = 3.91, specific conductance
= 50/yScm"1 at STP).

Hydrolab measurements were taken at the depth
from which the  sample was collected,  which was
usually 1.5 m. A second set of readings was taken
at 1.5  m above the  lake bottom. If the temperature
difference between the two measurements taken at
the top and bottom was <4°C, the lake was classified
as nonstratified. If the temperature  difference
between the two depths was > 4°C, a  third set of
readings 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, and vertical profiles of temperature and
conductance were determined.

Ground crews collected water samples  and  took in
situ measurements  from inflatable boats. Because
of logistical constraints, ground crews did  not use
Hydrolab units; therefore the protocol for taking in
situ measurements  differed from that used by the
helicopter crews. A Yellow Springs Instruments (YSI)
model  425C telethermometer with a YSI 400 series
probe was used to determine water temperature and
to check the stratification status as described above.
The telethermometer was checked daily against an
NBS-traceable thermometer, and was  recalibrated
if the temperature difference was more than ±0.5°C.
Site  depth and Secchi disk transparency were
determined as described for the helicopter sampling
procedures.

Water  samples were  collected  by helicopter and
ground crews with a 6.2-L Van Dorn acrylic plastic
sample bottle. Samples were collected from 1.5 m
if the site was deep enough to allow a clean sample
to be taken (i.e., a sample free from sediment, plants,
and other large  particles associated with the lake
bottom). If a clean sample could not be obtained from
1.5 m, the sample  was collected  from 0.75 m. If
a clean sample could not be collected from 0.75 m,
the lake was designated as a shallow lake and was
classified as non-target.

The Van Dorn sampler was modified to accommodate
a nylon  Luer-Lock  fitting  to  permit  collection  of
sample aliquots in  syringes without atmospheric
contact.  Two  60-mL  polyethylene syringes were
                                                                        77

-------
Figure 2-5.     Field sampling activities. Western Lake Survey - Phase I. Field crews accessed most sampling sites in wilderness
               areas by ground and in non-wilderness areas by helicopter.
                                        Hike/Pack
                                         to Lake
                       Fly to
                       Lake
                                             Verify Identity
                                        Record Site Characteristics
                                                     Photograph Lake
                                          Select Sampling Site
                               Flow Out and Anchor
                                                     Land on Lake
                         Determine
                        Stratification
                           Status
                     (Temperature Profile)
                        Other In Situ
                      Measurements
                   Secchi Disk Transparency
                   pH (Paper)
                 Record Data
                  in Logbook
 pH (Probe)
Conductance
                                 \ Return to Shore
Collect Water Samples
Field
Blank
Sample
(Deionized
water)
4-L
Container

125-mL
Bottle
Routine
Sample
(from 1.5 m)
4-L
Container
Two 60-mL
Syringes
125-mL
Bottle
Field
Duplicate
Sample
(from 1.5 mj
4-L
Container
Two 60-mL
Syringes
125-mL
Bottle
                                           Preserve 125-mL
                                          Aliquots with HgC/2
                                             Complete Field
                                              Data Form
                                       Hike/Pack to
                                        Trailhead
                         Fly to
                       Next Lake
                                           Deliver Samples to
                                            Field Laboratory
                            18

-------
attached, in turn, to the fitting, were filled with water,
and were sealed with locking syringe valves. These
syringes were used for pH and dissolved inorganic
carbon (DIC) analyses at the field laboratory. A bulk
water sample was collected by completely filling a
4-L polyethylene Cubitainer from  the Van  Dorn
sampler. The syringes and Cubitainers were placed
in coolers with frozen chemical refrigerant packs and
were delivered  to the  field  laboratory.  Although
helicopter crews and ground crews both followed
this protocol for sample collection, ground crews also
collected a 125-mLaliquot in an opaque polyethylene
bottle. This aliquot, which was preserved with 0.1
mL of 5 percent HgCl2 at the lake  site, was used
as  an independent check of the chemical stability
of nitrate and sulfate.

Two types  of QA samples (Section 3.1.5.1), field
blanks and field duplicates, were collected by the
sampling crews. Field blank samples were  "col-
lected"  by filling the  Van  Dorn  sampler  with
deionized water (obtained from the field laboratory)
which was then used to fill a 4-L Cubitainer. Because
deionized water is especially prone to CO2 contam-
ination as a result of  atmospheric contact, syringe
samples for pH and DIC analyses were  not collected
from field blanks. Field duplicate samples, consisting
of  two  syringes and one 4-L  Cubitainer, were
collected by obtaining a  second Van Dorn sample
from a lake, at the same location and  depth as the
routine sample.

2.5.2  Field Laboratory Operations

Five field laboratory trailers, one at each field station,
provided facilities  for  sample receipt, analysis,
aliquot preparation, preservation, and shipping. The
field laboratory also facilitated distribution of water
to  the  field  crews  for blank samples and of
conductivity standards for Hydrolab calibration. Field
laboratory  personnel  included a laboratory coordi-
nator, a  laboratory supervisor, and  three analysts.
Laboratory activities (Figure 2-6) were  conducted in
accordance with established protocols for laboratory
safety and  maintenance and included preparation
of  reagents,  blanks, and  samples;  instrument
calibration; data  recording; and supply inventory.
Morris et al. (1986) give a detailed discussion of field
laboratory protocols.
The field laboratory coordinator, who  had  overall
responsibility for  administration of the field  labor-
atory, provided the helicopter crews with water  for
blank samples and  with conductivity  standards.
Supplies for the ground crews were provided through
the field laboratory coordinator  by the logistics
coordinator. The laboratory coordinator received lake
samples, lake data  forms,  and chain-of-custody
forms (from the  ground crews through the Forest
Service  field   manager)   and  received field  audit
samples from a designated supplier (Radian Corpo-
ration, Austin,  Texas). The laboratory coordinator
organized  the  lake samples,  blanks,  and  audit
samples into "batches," i.e., all samples processed
at one field laboratory on a single day. Samples were
then  assigned  batch and  sample  ID numbers  to
facilitate sample tracking. After the sample batches
were organized, the laboratory coordinator refriger-
ated the samples at 4°C until they were removed
for analysis or processing. The laboratory coordinator
also was responsible for maintaining and procuring
all supplies.

The laboratory supervisor, who was responsible for
overseeing daily  laboratory operations, performed
DIC and pH determinations. These two measure-
ments were conducted on the syringe samples, and
are termed "closed system" measurements because
the samples did  not come into  contact with the
atmosphere  during collection or analysis.  The
supervisor also designated one sample in each batch
as the trailer duplicate (Section 3.1.5.2). This sample
was analyzed in duplicate for DIC, pH, turbidity, and
true color.

Three laboratory analysts shared sample processing
and sample analysis duties. One analyst performed
aluminum extractions, the second performed sample
filtration,  and  the  third prepared  and preserved
aliquots and split samples (Table 2-6). Split samples
were prepared  from routine lake samples  both  in
the field laboratory and by field sampling crews for
special studies conducted at the EPA Environmental
Research  Laboratory in Corvallis,  Oregon (ERL-
Corvallis) and  the EPA Environmental Monitoring
Systems Laboratory  in Las Vegas, Nevada (EMSL-
Las Vegas). The third analyst also performed turbidity
and true  color determinations. All laboratory
personnel were responsiblefor maintaining logbooks
of their respective tasks; these records were checked
daily by the laboratory supervisor.

After a set of aliquots was processed, each container
was sealed with electrical  tape and was placed  in
an  individual plastic bag which was then  sealed.
Next, each set of aliquot containers was placed  in
a one-gallon Ziploc  bag,  which was also  sealed.
Containers were refrigerated at 4°C until shipment.
For transport to the analytical laboratories, each set
of aliquots and the corresponding shipping forms
were placed in  an insulated shipping box that was
maintained at  approximately 4°C. The boxes were
shipped to  the analytical laboratories via overnight
courier service.

All  data and shipping  forms that were completed
at the lake site and at the field laboratory were
reviewed by the laboratory coordinator. Copies of all
forms were then  sent  to EMSL-Las Vegas, where
                                                                          19

-------
Figure 2-6.    Field laboratory activities. Western Lake Survey - Phase I. Samples collected by field crews were processed
              at mobile field laboratories before shipment to analytical laboratories for analysis.
Sample Transfer Supply Requests from


*
Laboratory

1
DIC
Determination





,

'
Laboratory
Coordinator
,
1
'
Check in Audit Samples
and Organize Batch


Ana
3
,
i
lyst Analyst Analyst Logit


i
tics
nator
,111
Aliquot
Preparation
and
Preservation
pH
Determination

L

Turbidity
etermination



True Color
Determination

Data
Transfer



Sample Aluminum Track In
Filtration Extractions for Cr
Samolin



Shipping
Preparation










|

ventory
ound
g Crews

Field Supplies


Prepare
Reagents for
Field Crews

*
Pack Aliquots
Shipping Form ___
* uata horrns
and Supplies

                     Ship Aliquots
                    and Data Forms
 Transfer Supplies
to Sampling Crews
                           20

-------
they were reviewed for data consistency (Section
3.2.1). Copies of data forms were also sent to the
data base manager at Oak Ridge National Laboratory
(ORNL) in Oak Ridge, Tennessee, for use in data entry
(Section  2.7).

2.5.3  Analytical Laboratory Operations

Analytical laboratory personnel were responsible for
receiving the samples shipped to them  by the field
laboratory,  inspecting the samples for  damage,
logging in the sample batches, analyzing the samples
according to procedures described in the statement
of work and the WLS-I analytical  methods manual
(Kerfoot  and Faber  1987), and  preparing  and
distributing  data  packages on the  analyses  per-
formed.  Analytical methods are described  in
Volume II.

Each sample received by the analytical laboratory
consisted of the seven aliquots prepared at the field
laboratory. Each aliquot was processed in a different
manner depending on the analytes to  be  analyzed
(Table  2-6). In addition, each analyte had to be
measured within a specified holding time (Table 2-7).

For each batch  of samples,  analytical laboratory
personnel  completed a  data  package. The  data
package consisted of analytical results and support-
ing  information  from  the field and analytical
laboratories (Silverstein  et al. 1987a). On the basis
of the  analytical results reported for internal and
external  QA/QC  samples (Section  3.1.5) the QA
staff, with the approval of the QA manager, could
direct the analytical laboratory to  confirm reported
values or to reanalyze selected samples or sample
batches.   In such  cases,  sample reanalysis  was
usually requested for a  given parameter on a per-
batch basis.

2.6  Calibration Study

The  standardized protocol  for the  National  Lake
Survey prescribed sampling by helicopter; however,
more than half the lakes selected for  sampling in
the WLS-I (Section 2.2.3) were  located in  Forest
Service wilderness areas, where access by moto-
rized vehicles is prohibited (Wilderness Act of 1 964).
Access to these  sampling sites  by ground  was
expected to  be more difficult than was access by
helicopter. Access to the sampling site, identification
of the lake,  and location of the correct  site on the
lake for sampling are more readily accomplished by
helicopter. Additionally,  adherence to  established
sample handling protocols, particularly delivering
samples  to  the field laboratory within prescribed
holding times, was expected to be more difficult for
ground crews than for helicopter crews.
To  evaluate the differences between  the two
sampling  methods,  the Forest Service granted
permission to perform a "calibration"  study on  a
subset of the lakes located in wilderness areas. Each
calibration lake was  sampled by a  helicopter crew
and a ground crew. Results of analyses of samples
collected using these methods  were compared to
determine whether or not calibration of the chemical
variables was necessary. Specifically, the data were
compared  to evaluate the practical and statistical
significance  of potential  differences  between  the
sampling protocols and the effect of holding samples
for  different lengths of time before  processing,
preservation, and analysis.  Additionally, samples
collected as  part of the calibration study were used
to evaluate potential bias between the two analytical
laboratories. The results of these analyses are given
in Section 4.3 and in the Appendix.

The 50 calibration  study lakes were selected as  a
systematic random sample from a list of the Forest
Service wilderness area  lakes ordered by lake ID
(Section 2.2.5). This procedure  permitted  the
calibration study lakes to be distributed proportion-
ately among subregions and wilderness areas. Five
lakes originally determined notto be in Forest Service
wilderness areas (and thus not included in the pool
from which the calibration study sample was drawn)
were later found to be in Forest Service wilderness
areas  and were sampled  by ground crews. In
addition, six lakes in national parks and another lake,
to which  access by  helicopter  was denied, were
sampled by  ground access;  these  lakes were not
included  in  the original pool  from  which  the
calibration study sample was drawn.

Both  crews (helicopter and ground) collected
samples from  approximately the same  location on
the lake. When logistically possible, the ground crew
sampled the lake  first, and  the helicopter crew
sampled the lake as soon as possible thereafter. Of
the 45  lakes sampled in the  calibration study, 30
lakes were sampled by both crews on the same day.
Twenty-three of the 30 lakes  were  sampled by the
ground crews first and seven of the 30 lakes were
sampled by  the  helicopter crews first.  The period
of time between sampling for  these  lakes  never
exceeded five hours.  The remaining 15 lakes were
sampled on different days, in most cases as a result
of weather-related delays. The ground crew collected
a routine and a duplicate sample; the helicopter crew
collected a  routine,  a duplicate, and  a triplicate
sample.

To  ensure  that each ground crew's sampling
procedure would not  be biased by knowledge of the
calibration lakes, the ground  crews were not told
which  lakes were part of the study.  Because the
helicopter  sample collection procedure  had been  a
                                                                        27

-------
Table 2-6.    Aliquots and Corresponding Measured Parameters, Western Lake Survey - Phase I
          Aliquot'
                     Container
                    Volume (mL)
      Preservative and
         Description
          Variable
Field Laboratory Aliquots
  Syringe A
  Syringe B
  Cubitainer
  Cubitamer

Analytical Laboratory Aliquots
  1
  2
  3
  4
  5
  6
  7

Split Sample Aliquots
  EMSL-Las Vegas Splits
  ERL-Corvallis Splits
                        60
                        60
                        25
                        50
                       250
                        10
                       250
                       125
                       500
                       125
                       125
                       125
                       125
Unfiltered
Unfiltered
Centrifuged
Unfiltered
Filtered, pH < 2 with HN03
Filtered, MIBK-HO" extract
Filtered
Filtered, pH < 2 with H2SO4
Unfiltered
Unfiltered, pH < 2 with H2SO4
Unfiltered, pH < 2 with HN03
Unfiltered, 0.1 mL 5% HgCI2
Filtered, pH < 2 with HN03
pH, closed system
DIG, closed system
True color
Turbidity
Ca, Fe, K, Mg, Mn, Na
Extractable At
cr, r, NO3", Sio2, so4~2
DOC, NH/
ANC, DIG, pH, conductance
Total P
Total Al
NO3~, S04"2
Metals, SCU'2, Si02
"All ahquots and the EMSL-Las Vegas split samples were stored at 4°C in darkness.
 ERL-Corvallis split samples were stored at room temperature.
"Methyl-isobutyl ketone, hydroxyquinolme.
Table 2-7.   Maximum Holding Times Specified for Samples,
           Western Lake Survey - Phase I
Holding Time
                        Parameter
7 days
14 days
28 days

6 months3
N03  , air-equilibrated pH, extractable Al
ANC, conductance, DIG, DOC
Total P, NH4 + ,  Cl~, SO4~2, total dissolved
  F~, Si02
Ca, Fe, K, Mg,  Mn, Na, total Al
a Although holding time has been established at 6 months (U.S.
 EPA 1983), samples were required to be analyzed within 28
 days  for these parameters  due to survey data reporting
 restrictions (Linthurst et al. 1986).
tested and proven protocol in the ELS-I, there was
no need for the helicopter crews to be unaware of
the identity of calibration lakes.

Concern that the ground sampling protocol might
result in delayed delivery of samples to the  field
laboratories for processing led to the  development
of an  experiment  using  three preservation  proce-
dures. Both of the ground crew's samples and two
of the helicopter crew's samples were preserved at
the field laboratory on the day of collection.  Of the
three samples collected by the helicopter crew, one
was selected randomly and  was  held in the  dark
at 4°C at the field laboratory for a specified period
of time before  it was processed and preserved. This
holding time depended on which of the three possible
processing procedures applied (Section 4.4). Further
discussion of the processing  and preservation-time
procedures for samples from the calibration lakes
appears in the WLS-I QA plan (Silverstein et  al.
1987a).
  To evaluate  interlaboratory bias,  all  samples
  collected for the calibration study were assigned to
  the two analytical laboratories for analyses. A die
  roll was  used to determine the assignment of the
  samples collected by the helicopter crews, and a coin
  toss was used to determine the  assignment of the
  sample collected by the  ground crews. This proce-
  dure (Figure 2-7) ensured that one sample collected
  by each  field crew (helicopter and ground) would
  be sent to each analytical laboratory.

  2.7   Data Base Management
  The objectives of the  data management component
  of the WLS-I were to enter the data,  prepare a
  validated data set,  and perform statistical analysis
  and  evaluation of the  data.  The Environmental
  Sciences  Division  of  ORNL designed  and imple-
  mented data base management activities using the
  Statistical  Analysis System (SAS) software  (SAS
  Institute  1982)  on tandem  IBM  3033  mainframe
  computers.

  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
  definitions of all variables measured,  their formats,
  units of  measure, and comments are presented in
  a  data  base dictionary  (Kanciruk et  al. 1986b).
  Detailed  discussions of the QA measures developed
  for the data  base are  given in Rosen  and Kanciruk
  (1985) and Kanciruk et al. (1986c).

  The  lake  ID assigned  before  field  work began
  (Section  2.2.5) was used as the key  site identifier.
                        22

-------
Figure 2-7.    Allocation of samples to analytical laboratories for the calibration study. Western Lake Survey - Phase I. This
             procedure was used to assign routine (R), duplicate (D), and triplicate (T) samples collected by ground crews
             (GC) and helicopter crews (HC) to one of two analytical laboratories.
Ground Samples
(Forest Service)
Routine
(1 st Sample
Taken)

Duplicate
(2nd Sample
Taken)
                                                              Helicopter Samples
                                                            {EPA, Lockheed-EMSCO)
     V
           RGC
                             DGC
                    V
                                      .
Routine
(1st Sample
Taken)
. RHC

Duplicate
(2nd Sample
Taken)
DHC

Triplicate
(3rd Sample
Taken)
THC
y
V
              Laboratory
                  1
                                       Randomly
                                       Selected
                                        Sample
                                       Shipment
                                      \    X
                                         X
                                        Withheld
                                       Helicopter
                                        Sample
                                       /
                                                 /
                                                \
          Laboratory
             2
The lake 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 assurance
procedures were implemented (Section  3). The final
reported data set (Data  Set 4) was developed from
                                                                             23

-------
Figure 2-8.     Procedures used to develop Data Set 1 (raw). Data Set 2 (verified). Data Set 3 (validated), and Data Set 4
               (final) for the Western Lake Survey - Phase I data base.
                       Visual Form Check
                   I
            Data Entry 1
 I
Data Entry 2
                        Error and
                        Range Check
                      Raw Data Set
                      (Data Set 1)
           Substitution
               and
           Replacement
              Final Data Set
              (Data Set 4)
                                                 Batch Reports
                                                      Verified Data Set
                                                         /Data Set 2)
                                                                              Verification
                                                                              Data Editing
                                                                              and Flagging
                                                                                Site
                                                                               Reports
                                                                                Maps
                                                                                                 Validation
                                                                                               Data Editing
                                                                                               and Flagging
                                                                                                    1
                                       Validated Data Set
                                       (Data Set 3)
                           24

-------
the validated data set by removing erroneous data
and substituting for missing values when appropriate
(Section 3.4). Protocols used for these changes are
discussed in Eilers et al.  (1987). Copies of all data
sets have been maintained on tape as a permanent
record.

Background information (lake names, lake IDs, and
physical characteristics) was transmitted from ERL-
Corvallis to ORNL, where it was entered into a data
file.  Data forms that contained 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 by EMSL-Las Vegas personnel.
Corrections and flags were returned to ORNL as the
verified data set. After verification, the data were
validated jointly by ERL-Corvallis, EMSL-Las Vegas,
and ORNL staff.
                                                                         25

-------
                                         Section 3
                                    Quality Assurance
3.1   Design of the Quality Assurance
Program

The WLS-I quality assurance (QA) and quality control
(QC) activities summarized  in  this section are
described in detail in Silverstein et al. (1987a). Most
of the QA/QC procedures were previously used in
the ELS-I (Drouse et al. 1986).

The QA/QC activities included (1) selecting analyt-
ical laboratories, (2) training field sampling and field
laboratory crews,  (3)  maintaining  communication
with management, sampling, and analytical person-
nel,  (4) conducting on-site field  and laboratory
inspections, and (5) collecting and analyzing a variety
of QA/QC  data to quantify data quality. Quality
assurance protocols were developed for collection,
preparation, preservation, shipment, and analysis of
samples, as well as for reporting, verification, and
validation of analytical results.

3.1.1  Selection of Analytical Laboratories

The objective of the selection process for analytical
laboratories was to choose the minimum number
of qualified laboratories needed  to analyze WLS-I
samples. The  Contract Laboratory Program estab-
lished to support EPA's hazardous waste monitoring
activities was the most  efficient  mechanism  for
accomplishing this objective.

The selection procedure was initiated by preparing
a statement of work that defined the analytical and
QA/QC requirements in a contractual format. Next,
bids were solicited from analytical laboratories. On
the basis of pre-award sample analyses and on-site
evaluations. Environmental  Monitoring  and Ser-
vices, Inc. (EMSI) in Thousand Oaks, California, and
Versar,  Inc., in Springfield,  Virginia, were selected
from among the responding  laboratories. During the
WLS-I, one laboratory processed about 40 percent
of the samples, primarily  from the Central (4D) and
Southern (4E)  Rockies, and the  other  laboratory
processed the remaining 60 percent (Table 3-1).

3.1.2   Training of Sampling and Field Laboratory
Crews

Because data quality  is  affected by the  ability of
personnel to collect, process, and analyze samples
properly,  training of the  WLS-I  personnel was
essential to  ensure consistent  application  of
sampling  and  analytical  protocols.  Before field
sampling activities began, field sampling and field
laboratory personnel from Lockheed-EMSCO were
trained in sampling and field laboratory procedures,
respectively. These  personnel, most of whom had
also participated in the ELS-I activities (Linthurst et
al. 1986), then were sent to the field stations where
they trained EPA and Forest Service personnel. This
ensured that field activities were performed consist-
ently and according to approved procedures.

3.1.3  Communications

Monitoring  of  QA  activities  required continuous
communication among the many groups responsible
for data  collection. During  the sampling and
analytical  phases,  these  communications were
centralized through the QA staff at EMSL-Las Vegas
who made daily calls to the field stations and to the
analytical laboratories. These calls were  used (1) to
determine whether  QA/QC guidelines were being
followed,  (2) to  check  that samples were being
processed and analyzed properly, (3) to obtain current
sample and QA/QC data, and (4) to discuss sampling,
processing, and analysis issues related to logistics,
methods,  and QA/QC so  that problems could be
resolved quickly and efficiently. All communications
were logged on appropriate field communications
forms and in bound notebooks.
3.1.4  On-Site Inspections

On-site inspections of field stations, remote sites,
and analytical laboratories were conducted during
the WLS-I to ensure that sampling and analytical
activities were  being performed according  to
established  protocols. The results of these  evalua-
tions were documented in site-evaluation  reports
prepared by the QA staff and presented to EPA.

3.1.5  Quality Assurance/Quality Control Data
Collection and Analysis

Specified  QA/QC procedures  and samples were
used in the field, at the field  laboratories,  and at
the analytical laboratories (Silverstein  et al.  1987a)
to maintain  the quality of the  data and  to  ensure
that data quality could be characterized accurately.
Rigid requirements for instrument calibration helped
                                               26

-------
Table 3-1.   Number of Samples from Probability Sample Lakes Analyzed by the Analytical Laboratories, by Subregion,
          Western Lake Survey - Phase I
Subregion
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
Non-Calibration Lakes

4A
4B
4C
4D
4E
Lab 1
6
0
0
115
130
Lab 2
135
149
133
7
0
Non-
Calibration
141
149
133
122
130
Total
Calibration3
9
10
10
7
9

Total
150
159
143
129
139
West
                               251
                                              424
                                                             675
                                                                             45
                                                                                           720
a Analyzed by both laboratories.
to provide reliable measurements. Data derived from
the  QA  and QC samples  were compared to the
expected  values  and established  ranges (Section
2.1.1, Table 2-1).

Several types of QA and QC samples (Figure  3-1)
were used to ensure that sampling and  analytical
methods were performed according to specifications.
The results of QA  sample analysis were used to
evaluate the performance of field sampling methods
and field  and analytical laboratory procedures and
to  evaluate  overall data  quality  for the survey.
Analysis of the QC  samples allowed field samplers
and laboratory personnel (in  field and  analytical
laboratories) to identify and correct immediately local
problems such as poor instrument performance or
reagent contamination before and during routine
sample analyses.

3.1.5.1   Quality Assurance  Samples— Quality
assurance samples, introduced into batches of  lake
samples  at the field station, were  analyzed at the
field laboratory and at the analytical laboratory.
Analytical laboratory personnel were not told the
origin, identity, or  chemical  composition of  the
samples. Thus, the  QA samples were analyzed as
if they were routine lake samples. The QA samples
used were  field  blanks, trailer blanks, field
duplicates, and field audits (Figure 3-1).

Field blanks were prepared at the field laboratory
from American Society for Testing and  Materials
(ASTM) Type 1  reagent-grade deionized water.  The
sampling  crew  transported  the blank  water in
Cubitainers to the sampling site and processed the
blank through the Van Dorn sampler as if the blank
were a lake sample. Each helicopter crew collected
one field blank on each operating day; each ground
crew collected two  field blanks during the course
of the survey.

Field blanks were added to the sample batches that
were sent to the analytical laboratories. These QA
samples  were  used  to  identify   potential
contamination  resulting  from  sampling and
analytical processes and to provide data to estimate
the  system  decision  and  detection  limits,
quantitation limit,  and background concentrations
(Sections 4.2 and 4.3) for each  variable. Measured
sample values at or below  these system decision
and  detection limits were considered to be indis-
tinguishable from blanks.

Occasionally,  due to  logistical constraints, field
blanks were not scheduled to be processed at any
site on a particular sampling  day. In these instances
a deionized water  sample,  prepared in the field
laboratory  without processing through the Van
Dorn sampler, was substituted  for the field  blank.
This "trailer blank" was then added to the sample
batch.

Field duplicates  were  collected immediately after
routine samples were collected. The same procedure
was used to collect both types of samples. For each
base site, one helicopter crew collected one field
duplicate on each sampling day. Each ground crew
collected two field duplicates during  the course of
their sampling. Field duplicates were processed by
the field laboratory and were added to the sample
batches  sent to the analytical laboratories.  Field
routine/duplicate pairs were used to determine the
homogeneity of the lake sample and the precision
of the  field sampling techniques, field laboratory
procedures, and analytical  laboratory  procedures
(Section 4.3).

Two types of field  audit samples (natural and low-
concentration synthetic) were prepared by Radian
Corporation in Austin,  Texas,  and were used to
evaluate overall  field  and  analytical  laboratory
performance  (Section 4.3).  A third type  of  audit
sample, a laboratory audit, was used during the ELS-I
(Linthurst et  al.  1986)  but  was  not used during
the WLS-I. An evaluation of data collected during
the ELS-I indicated that field audit results were no
more  variable than  laboratory audit  results;
therefore, the laboratory audits were deleted.
                                                                        27

-------
Figure 3-1.   Origin, use, and flow of quality assurance/quality control samples. Western Lake Survey - Phase I. Results
             of the analyses of these samples were used to quantify data quality and to ensure that error was minimized
             at each level of sample handling and processing.
Field
Samplers

I
Ftf\M P/rtnlr

Field
Laboratory

-" I


Analytical
Laboratory


ta. rVn/r/n/ln<

                                                      Trailer Blank
                                                  (in Lieu of Field Blank}
                                                 Trailer Blank
                     Field Duplicate
                Field Duplicate
 Field Duplicate
                                                     Trailer Duplicate
                                                   (Split of a Randomly
                                                     Selected Routine
                                                     Lake Sample)
                         QCCS
                    Hydrolab pH. Cond.
         Natural
         Audits
    Lake Superior, MN
         (FN3)
    Big Moose Lake, NY
          IFN4)
     BagleyLake, WA
       (FNS; FN6)
 Radian Corp.
Synthetic Audits
 (FL11;FL12)
 Prepared FNs
                 Field Audits
                                                        QCCS
                                                    pH, DIC, Turbidity
               Calibration Blank
                    DIC
                                                 Field Audits
                                                                                        QCCS
                                                 Calibration/
                                                Reagent Blank
  Matrix Spike
(on Field Sample)
                                                                                   Laboratory Duplicate
                                                                                  (Split of Field Sample)
 Field  natural  audit  samples were composed of
 natural lake water. Field  synthetic audit samples,
 which were prepared to simulate natural lake water,
 included a matrix of analytes at specified theoretical
 concentrations.  The  mean  concentrations of the
 audit samples are  given in Section 4.3. Field  audit
 samples were used  (1) to determine  relative  bias
 between   analytical   laboratories  so  that
 measurements made by the two laboratories could
 be compared and (2) to indicate precision  of those
 measurements through  repeated analysis of the
 same sample type.

 Both types of field audit  samples, received at the
 field laboratory  in 2-L aliquots, were filtered and
 processed into aliquots as if they were routine lake
 samples. The samples then were shipped to the
 analytical laboratory as if they were routine samples.
                i.e.,  their  identities  and  concentrations were
                unknown. Four field  natural  samples  (designated
                FNS, FN4, FN5, and FN6) and one low-concentration
                field synthetic sample (two lots, designated FL11 and
                FL12) were used.

                Three of the natural  samples (FN4 from Big Moose
                Lake in the Adirondack Mountains of New York, and
                two separate samples,  FN5 and FNS, from Bagley
                Lake inthe North Cascade Mountains of Washington)
                represented two  types  of low ANC and low ionic
                strength surface waters expected to be encountered
                during the  survey. The  fourth natural sample (FNS
                from Lake Superior) had relatively high ANC and high
                ionic strength.

                Following collection, each natural audit sample  lot
                was filtered  in bulk  and divided into  2-L aliquots
                       28

-------
(as individual audit samples), and the aliquots were
stored at 4°C until use. Field synthetic audit samples
were prepared as lot concentrates and diluted just
before they were  sent in 2-L bottles to the field
laboratory.

3.1.5.2  Field  Sampling  and  Field Laboratory
Quality Control Samples—Quality control check
samples (QCCSs) were used by the helicopter crews
to check the calibration  of  the  Hydrolab pH and
conductance functions before sampling and to check
for instrument drift after sampling. Daily QC checks
were made before and  after sampling (Section
2.5.1.2).

Three types of QC samples were used by the field
laboratory staff to ensure that instruments and data
collection  were  within  specified  control limits
(Silverstein et al.  1987a). A  calibration  blank,
analyzed before samples, was  used to check for
baseline drift of the carbon analyzer and to check
for contamination. Calibration and drift of the carbon
analyzer and of  the instruments used to measure
pH and turbidity were checked with QCCSs. A "trailer
duplicate" sample was used to evaluate the precision
of  field laboratory  measurements. The field
laboratory supervisor  randomly selected  one lake
sample  per  operating  day as the trailer duplicate,
which was analyzed in duplicate for pH,  DIG, true
color, and  turbidity.

3.1.5.3 Analytical Laboratory  Quality Control
Samples—Six types of QC  samples were used  by
the analytical laboratories to ensure that instruments
and data collection were within control limits. These
six QC sample types were (1) calibration blanks, (2)
reagent blanks, (3) detection limit QCCSs, (4) low-
concentration  and high-concentration QCCSs, (5)
matrix spikes, and (6) laboratory duplicates. The use
of these QC samples  in evaluating data quality is
described in detail in Silverstein et al. (1987a).

3.2  Data Verification

The objectives of the data verification process were
to identify, correct, and "flag" raw data of question-
able and unacceptable quality, and to identify data
that may have had to be eliminated during or after
validation  (Section 3.3). These objectives were
accomplished by reviewing  QA/QC data  recorded
at the sampling site, at the  field  laboratory, and at
the analytical laboratory. Computer programs were
used to automate the verification procedure as much
as possible. For each data package (representing one
sample  batch) the audit team performed a sample-
by-sample  evaluation.  The  audit  team  reviewed
comments and questions associated with the sample
batch; performed QA checks for data consistency and
reasonableness; reviewed QA sample data; obtained
confirmation, correction, and  reanalysis data from
the analytical  laboratories; and provided a verified
data set to ORNL. For each batch, the audit team
prepared a summary of the reporting errors and the
required confirmation and reanalyses.

3.2.1   Review of Field Data Forms

Verification began with the receipt of the data forms
from the field. The auditor reviewed the field forms
for the following items: lake ID, instrument calibra-
tion data, pH and DIG sample and QCCS data, site
depth and Secchi disk transparency, Loran-C  and
map coordinates, and tags (data qualifiers assigned
by field  sampling  crews  and field  laboratory
personnel). Data anomalies were reported to the field
laboratory coordinator for corrective action, and data
reporting errors were corrected before the data were
entered into the raw data set.

3.2.2   Initial Review of Analytical Laboratory
Sample Data Package

The sample  data  packages received from  the
analytical laboratories were reviewed for complete-
ness, internal QC compliance, and appropriate use
of data qualifiers. A checklist was used by the auditor
to ensure consistency  in  the  review  of all  data
packages. Problems were reported to the analytical
laboratory manager for corrective action. Comments
that the  laboratory  had  submitted with  the  data
package cover letter also were reviewed to determine
their impact on  data quality and to determine any
need for follow-up action by the laboratory.

3.2.3   Review of Quality Assurance/Quality
Control Data

After the data were entered into the raw data  set,
a magnetic tape containing the data was sent to the
National Computer Center in Research Triangle Park,
North Carolina. The QA personnel  then were able
to access the data by telecommunication. A series
of computer programs that comprise the Aquarius
QA/QC system  (Fountain et al. 1986) was used
during  verification.  These programs were used to
identify  or flag   results  that were classified  as
"exceptions"  (i.e., results  that did not meet  the
expected QA/QC limits). The auditor used the output
from these programs (with original raw data and field
notebooks) to complete the verification  report form
specified in the QA plan  (Silverstein et al. 1987a).
The verification report was  a worksheet designed
to guide  the  auditor systematically through  the
verification process by explaining howto(1)flag data,
(2) track  data resubmissions and requests for
reanalysis and confirmation, (3) list the steps that
lead  to identification of QA exceptions, and (4)
summarize modifications to the raw data set  (i.e.,
to prepare records of numeric and flag changes).
                                                                        29

-------
Each sample was  verified individually  and  by
analytical batch. The sample had to meet checks for
anion/cation percent ion balance difference and for
percent difference  between calculated and  mea-
sured conductance in order not to be flagged, unless
the discrepancy could be explained by either the
presence  of organic species (as indicated  by  a
protolyte analysis  program) or by an obvious and
correctable  reporting error. The ion  balance was
determined by converting measured values in mg L~1
to fjeq L~\ and then comparing the sum of measured
anions  to   the  sum  of  measured  cations.
Conversion factors are given in Table 3-2. Calculated
conductance was determined by using the conver-
sion factors in Table 3-3. Additional data qualifiers
were added to a given variable when the QA samples
(field blanks, field duplicates,  or audit  samples)
within the same analytical batch did  not  meet the
acceptance  criteria. Data  were also qualified  if
internal QC checks such as matrix spike  recovery,
calibration  and  reagent blank  analyses,  internal
duplicate precision, required instrumental detection
limit, QCCS percent  recovery, and required holding
times were not met.  The protolyte analysis program
flagged field laboratory and analytical laboratory
measurements of pH, DIG, ANC, and  DOC (Drouse
et al. 1986). In  all cases,  each flag  generated  by
the Aquarius system was evaluated by the auditor
for  reasonableness and consistency before it was
entered into the verified data  set. These programs
Table 3-2.
Ion
           Factors Used to Convert mg L~1 to ^ieq L~1 for
           Anion/Cation Balance Check8, Western Lake
           Survey - Phase I
            Factor
                              Ion
                                          Factor
Ca + 2
cr
Mg + 2
N03-
F-
49.9
28.2
82.3
16.1
52.6
NH4+
K +
Na +
so4-2

55.4
25.6
43.5
20.8

aHillman et al. (1986).
 Table 3-3.   Factors Used to Convert mg L  1 to pS cm 1 for
           Determining Calculated Conductance"-6, Western
           Lake Survey - Phase I
 Ion
      Conductance Factors0
                          Ion
                                Conductance Factors0
Ca+2
cr
co3-2
H +
HC03-
Mgt2
2.60
2.14
2.82
3.5 x 10B
0.715
3.82
Na+
NH4 +
SO4~2
N03-
K +
OH' 1
2.13
4.13
1.54
1.15
1.84
.92 x 105 (mole





L")
 aHillmanetal. (1986).
 b Factors for H+ and OH~ are used to convert mole L'1 to
      '1
 /iS cm
cConductance
 strength.
               cm"1) factors at 25°C for 0.001 ionic
also identified outlier data based on QA and QC
samples. The outlier  data were  the basis for
requesting confirmation of data from the analytical
laboratories  and  for  reanalysis  of  suspicious
samples. Data verification and data validation
(Section 3.3) were  performed in tandem;  however,
the verification process had to  be completed before
the validation process could be  completed (i.e., Data
Set 2  had to be in place before final conclusions
could be drawn from Data Set 3).

3.3   Data Validation

The  data validation  process  identified  possible
systematic and random errors in chemical analyses
that were  not  always  revealed by verification
procedures  (Eilers  et al.  1987a). The  validation
process  closely followed that  used in the  ELS-I
(Linthurst et al.  1986). The validity of nonchemical
measurements (e.g., lake area, watershed area, lake
depth and elevation) was also evaluated.

Potential random errors were  identified by gener-
ating a list of outliers, i.e., observations that were
not typical of other sample values.  Outliers were
detected using a variety of univariate, bivariate, and
multivariate approaches. Initially, each variable was
considered individually to identify values that were
outliers with respect to the sample distribution.

Box  plots (Tukey  1977)  for each  variable  were
prepared  using SAS statistical  procedures (SAS
Institute 1982). These plots summarized the data for
a variable by using the difference between the upper
and  lower quartile (interquartile  range). For this
procedure, outliers were defined arbitrarily 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 and least-squares linear
regression analyses.  Standardized residuals were
calculated as follows:
                                                             actual value - predicted value | /
                                                             residual standard deviation I
Values for which the standardized residuals were
greater  than  three  were  identified  arbitrarily as
outliers.  Because least-squares analysis can  be
biased strongly by  certain types of outliers,  the
residuals from resistant line fits (i.e., lines fit through
the medians  of partitions of data,  Velleman  and
Hoaglin 1981) were examined for selected variables
measured in the field and field laboratory. Analytical
variables were examined  using  a process of linear
regression with three iterative steps. Sequentially
removing outliers is a helpful technique for locating
additional  outliers that  would  not  have been
identified  without previously removing the  major
                       30

-------
aberrant values. Outliers among related groups of
variables were detected by using the SAS FASTCLUS
cluster analysis procedure (SAS Institute 1982) and
principal components analysis.

Statistical outliers are useful for highlighting suspect
data, but considerable caution must be exercised in
the  treatment of these outliers. It  cannot  be
presumed that all outliers  are  errors. Once the
outliers were identified,  it  was necessary to
determine the reason for the aberrant data.

Outliers were initially screened by comparing them
to other related variables that might explain  their
high residual standard deviation.  For  example,  a
sample  with a suspiciously high concentration of
nitrate might appear reasonable  if the sample also
exhibits  high concentrations  of other nutrients.
Those outliers  remaining  after  screening and
confirmation of the reported data were flagged in
the validated  data set. In a few cases, sufficient
evidence was available to indicate that the reported
value was erroneous. These values were flagged in
the validated set and were  marked for substitution
in the final data set (Section 3.4).

Analytical  variables  having possible  systematic
errors were  detected by comparing values from the
WLS-I to those from other lake survey data sources.
These  data  sets were selected on  the basis of
geographic location,  accessibility of the data, and
documentation of sampling and analytical methods.
Comparisons were  made using  scatter plots and
linear  regression  procedures to  identify  values
requiring additional scrutiny. This  approach failed
to identify systematic differences that  were suffi-
ciently large to indicate major analytical errors or
differences of a magnitude that  would affect data
analysis and interpretation. The details  of  the
validation process are described  in Eilers  et  al.
(1987a).

3.4  Development of Final Data Set

Calculating  population estimates (Section  2.3) is
difficult if values are missing from the data set. A
final data set (Data Set 4) was prepared to resolve
problems relative to data  interpretation because of
missing values in the validated data  set. Data Set
4 was modified by averaging field duplicate values
and substituting for analytical values determined to
be in error during validation.

Several substitution  methods  were  used.  Values
from duplicate samples were used when available.
Redundant analyses were  performed for pH, DIG, and
conductance (Section  2.5).  Redundant measure-
ments  on  split samples  (Section  2.5.2) were
performed  for  metals and  other  elements.  If  a
duplicate measurement was  not  available,  a
comparable measurement (e.g., in situ conductance
or analytical  laboratory conductance) was chosen
and was substituted for the missing  value using a
linear  regression routine.  If redundant measure-
ments were not available or acceptable, a substi-
tution value was calculated from the  available data
using observed relationships with other variables
(e.g., calcium and ANC). The last option used for
identifying  a  substitution  value  was to  use the
stratum mean. All substitution values were exam-
ined a second time for acceptability before including
them  in the final data set. Substituted values were
flagged as such in the final data set. Only 56 values
out of approximately 20,000 chemical measure-
ments examined during the validation process were
substituted in the final data set. Of these 56 values,
only 2 were missing from the data collection process.

Two other changes were made in the final data set.
If  duplicate data met QA  precision criteria, the
average value of the duplicates was used in the final
data set. Negative values for parameters other than
acid  neutralizing  capacity that resulted from
analytical calibration  bias were  set  equal to  zero.
The bias in the estimate of variance due to  this
adjustment did  not affect data analyses. All values
modified in the final data set were flagged to indicate
that they did not represent original  measurement
values.
                                                                         31

-------
                                            Section 4
                      Quality Assurance and Calibration Study Results
4.1   Site Confirmation

Confirmation that the lakes sampled by helicopter
were the lakes intended to be  sampled in the WLS-I
was accomplished by comparing aerial photographs
taken during the survey to lakes as they appeared
on the topographic maps. Lakes accessed by ground
crews were checked by comparing comments and
observed lake outlines noted on the field forms with
topographic  maps. Two  lakes  were  mistakenly
sampled. The data from these lakes were  deleted
from the final data  set and the  lakes targeted for
sampling were  classified as "not visited" (Section
2.2.2). Three lakes were determined after sampling
to be non-target; these data were also removed from
the data base. For two of these lakes, conductance
values exceeded the  maximum  criterion of  1500
/jS cm"1; the third  lake  was  determined to  be a
livestock watering pond.
4.2  Detection Limits (Detectability)

Helicopter crews  and ground crews collected 236
field blanks  during the WLS-I that were used for
statistical analyses. Data from the analysis of these
blanks were  used to estimate the overall background
contamination that occurred during sampling and
analysis. Two criteria (Table  4-1) were  used  to
evaluate analytical detectability: the system decision
limit (SDL) and the  required detection limit  (RDL).
The SDL for each variable is based on a nonpara-
metric statistical evaluation of  the  field blanks
(Section  3.1.5.1), and is  the concentration limit
(background) above which an analyte can be detected
with a known degree of confidence.  This limit  is
defined as the 95th percentile (P95) of the distribution
of field blank measurements. The SDL provides  an
estimate of the level of an analyte that  potentially
can be  introduced during sample collection, han-
dling, processing, or analysis. For measured values
below the SDL, it cannot be known with certainty
whether the analyte was present in the lake or was
introduced at some stage  of handling. The SDL  is
compared tothe RDL, i.e., the level of detection which
the analytical  laboratory was required by contract
to meet when analyzing laboratory blanks. Because
field blanks were subject to more handling than were
Table 4-1. Detectability Based on Evaluation of Field
Data8, Western Lake Survey - Phase 1
Variable13
Al, extractable
Al, total
ANC (Meq L"1)
Ca
cr
Conductance (^S cm"1)
DIG, air equilibrated
DIG, initial
DOC
F~, total dissolved
Fe
K
Mg
Mn
Na
IMH4 +
N03-
P, total
SiO2
so4-2
SDLC
0.004
0.019
3.9
0.07
0.04
1.6
0.33
0.43
0.3
0.003
0.01
0.01
0.01
0.01
0.01
0.01
0.071
0.006
0.18
0.07
Blank
RDLd
0.005
0.005
e
0.01
0.01
f
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
apH was not determined for field blanks and thus decision and
 detection limits do not apply to pH measurements.
bAII units are reported as mg L~1 unless otherwise indicated.
cSystem decision limit; the 95th percentile of 236 field blank
 measurements.
d Required detection limit.
e Absolute blank values were required to be < 10 /jeq I/1
f This value, the mean of six nonconsecutive blank analyses,
 was required not to exceed 0.9 j^S cm"1.

laboratory blanks, background concentrations infield
blanks were expected to be higher than levels in
laboratory blanks.

The quality assurance results are presented in  the
units reported by the analytical  laboratories (gener-
ally, in mg L"1). Throughout the remainder of  the
report, the  results are  presented  in  units  more
convenient for  direct comparison among variables
(generally pieq L"1). Conversion factors are given in
Section 3.2.3, Tables 3-2 and 3-3.

The  SDL was  below or near  the  RDL for  most
variables. For  total  Al,  Cl~, DIG (initial and  air
equilibrated), and  DOC, the SDLs were  notably
higher than the RDLs. However, the SDLs for these
parameters  are comparable to  field  blank data
collected in the ELS-I. Field blank data are discussed
in detail in Silverstein et al. (1987b). Other analytes
                                                32

-------
(Ca, NOa , and Si02) for which the SDLs were outside
the preferred range are discussed below.

The SDL for N0a~ was 0.071 mg  L"1. The concen-
trations of  NOs" in  WLS-I  lake samples were
generally very low. Therefore, nitrate was not a major
contributor of anion  equivalents in the  samples
collected from WLS-I lakes in the fall. As with N03~,
more  than  5 percent of the field blanks collected
had concentrations of Ca and Si02 that caused the
SDLs  to be substantial. The SDL for Ca was 0.07
mg L"1, and the SDL for Si02 was 0.18 mg L"1.
Therefore, there is a 5 percent, or greater, chance
of introducing contamination at levels above these
limits for Ca and Si02.

These results indicate that there was no evidence
for systematic contamination for  any  analyte.
However,  random contamination  as a  result of
sampling, processing, or analytical error may have
caused the SDLto exceed the RDLfor some analytes
(Section 4.6).
4.3  Precision

Sampling  and analytical  variance, apart from
seasonal variations in lake chemistry, can arise from
three major sources: (1) a field component associated
with short-term variability in lake chemistry, (2) an
analytical  component  associated  with aliquot
preparation or with variation in instrument response
within an  analytical batch,  and  (3)  an analytical
component associated with batch-to-batch variation
in instrument calibration and response. The relative
importance of these sources  of variation was
assessed by comparative statistical evaluations of
field audits, field  duplicates, trailer duplicates,  and
laboratory duplicates.

Precision can be calculated in two ways. One method
is by determining overall (among-batch) precision
using field  audit (natural and  synthetic) samples to
determine the precision across  batches  and over
time. The other method  is to determine  precision
within batches by using duplicate samples (field,
trailer,  and  analytical  laboratory)  to  estimate
precision associated with the different stages of
sample  handling  (i.e.,  sampling, processing,  and
analysis).

Overall among-batch precision for the WLS-I field
audits was calculated as  the standard deviation (SD)
and  as  the  percent  relative standard deviation
(%RSD) except for pH, which was calculated as the
SD only.  The %RSD is expressed  as 100 times the
ratio of the standard deviation to the mean (x):

                 %RSD= 100^
where x = the mean of all sample concentrations.

Tables 4-2  and 4-3 show the mean concentrations,
%RSDs, and SDs for synthetic  and natural audits,
respectively. Data from the two field synthetic lots
(FL11 and FL1 2) were combined, or pooled, because
the theoretical concentrations of these two lots  are
the same. The  overall precision for most variables
in the six lots of field audits (two pooled synthetic
and four natural) was reasonable, considering that
they were analyzed among many batches, except for
those variables that had  mean concentrations so
close to the instrument  detection limit  that  the
precision estimates have little meaning.  Data users
can compare these mean concentrations to the SDLs
(Section  4.2  and  Tables  4-2 and  4-3) to  gauge
whether the calculated precision is within acceptable
limits for a particular application. Among  all the field
audits, no precision estimates were consistently high
for any  one variable,  and  no specific trend of
imprecision was observed. A detailed, audit-by-audit
discussion  is not  within the scope of  this  report
because there are different concentrations for every
variable in each of the five audit  types (one synthetic
formula and four natural systems). Silverstein et al.
(1987b) provide this detailed discussion.

For  the duplicate samples (field,  trailer, and
laboratory), within-batch  precision  is expressed as
the root-mean-square (RMS) of the %RSDs of all
sample pairs, except for pH,  for  which  within-batch
precision is expressed as the RMS of the SDs of
the sample pairs.  To determine  if duplicate sample
pair concentrations  are sufficiently  above the
detection  limit to estimate precision  reliably, a
quantitation limit is calculated for all variables except
pH. The quantitation limit was calculated as  10 times
the  SD  (10  SB)   of the  concentrations of the
corresponding blanks (field, trailer, or laboratory).
Precision  estimates can  be calculated from  all
sample pairs (pairs with x > 0), some  of which  are
greatly affected by background (pairs with x near
0)  or from those pairs  minimally  affected  by
background (only pairs with x > 10 SB). Tables 4-4,
4-5, and 4-6 compare the values for all pairs with
x > 0 to only  those pairs with values  above the
quantitation limit (x > 10 SB).

Overall within-batch precision estimates calculated
from field duplicate pairs  are shown in  Table 4-4.
Like field blank SDLs, which were expected to be
higher than RDLs (Section 4.2),  field duplicate pairs
were expected  to  be  less precise  than laboratory
duplicate pairs. Field duplicates were  subjected to
more handling in less controlled environments (e.g.,
the lake sites) than were  laboratory duplicates. In
addition, field duplicates  were  truly separate
samples, whereas  laboratory duplicates were split
samples  of a  particular aliquot that  was created
                                                                         33

-------
Table 4-2.
Overall Among-Batch Precision Estimates Calculated from Field Synthetic Audit Samples, Western Lake Survey -
Phase I
Variable3
Al, extractable
Al, total
ANC (jtfsq L~1)
Ca
cr
Conductance (/^S cm"1)
DIG, air equilibrated
DIG, initial
DOC
F~, total dissolved
Fe
K
Mg
Mn
Na
NH4 +
NO3-
P, total
pH, acidity (pH units)
pH, alkalinity (pH units)
pH, air equilibrated (pH units)
Si02
SO4~2
Theoretical
Concentration
0.020
0.020
e
0.19
0.34
e
e
0.96
1.0
0.04
0.06
0.20
0.45
0.10
2.74
0.17
0.464
0.027
e
e
e
1.07
2.28
SDLb
0.004
0.019
3.9
0.07
0.04
1.6
0.33
0.43
0.3
0.003
0.01
0.01
0.01
0.01
0.01
0.01
0.07
0.006
g
g
g
0.18
0.07
Mean
Concentration0
0.005
0.027
111.0
0.22
0.36
19.7
1.44
1.54
1.0f
0.043
0.10
0.21
0.45
0.10
2.77
0.1 6f
0.483
0.025
6.94
6.95
7.24
1.10
2.30
Precision
% RSD
48. Od
29.3
6.1
16.1
5.6
4.4
18.3
20.4
25.0
7.1
153.0d
8.7
3.8
14.0
8.5
16.2
6.5
22.0
h
h
h
10.5
5.4
SD
0.002
0.008
6.76
0.03
0.02
0.87
0.26
0.31
0.26
0.003
0.009
0.02
0.02
0.01
0.24
0.03
0.03
0.005
013
0.11
0.14
0.12
0.12
aAII units are reported as mg L 1  unless otherwise indicated.
bSystem decision limit; the 95th percentile of 236 field blank measurements.
cMean concentration of pooled field synthetic lot 11 (FL11, n = 30) and field synthetic lot 12 (FL12, n = 17).
dPoor precision may be the result  of sample instability, sample mixing error, or both (Best et al. 1986).
eAlthough theoretical values for ANC, DIC, and pH can be calculated, the theoretical value is dependent upon the concentrations of
 chemicals added to the synthetic audit sample.
f n = 45.
9Decision and detection limits are not applicable to pH measurements.
hOnly standard deviation values are calculated for pH measurements.
within a controlled environment in the analytical
laboratory. Therefore, the DQOs for intralaboratory
precision  that were  used  to check analytical
laboratory precision should not be applied rigidly to
field  duplicate  precision. These DQOs,  however,
were useful as a  gauge to assess field  duplicate
precision estimates.

For field duplicates, eight variables met intralabor-
atory precision goals. These variables were total Al,
ANC, Ca, DIC (air-equilibrated and  initial), K, N03",
and  S04~2.  Seven  variables did   not  meet the
intralaboratory  precision goals, but  the  precision
estimates were  only  slightly higher  than the
intralaboratory precision goals. These variables were
CI",  Mg, Na, NH4+, pH (alkalinity and acidity), and
S\0z. Therefore, 15 of the 23 variables  presented
in Table 4-4 can be considered to have reasonable
to  excellent precision. For  one  variable (Mn),
precision above the quantitation limit could not  be
calculated because there was only one duplicate pair
where the mean of both sample concentrations was
greater than the quantitation limit.

Seven variables had poor field duplicate precision.
These were  extractable Al, conductance,  DOC, Fe,
total  dissolved F",  air-equilibrated pH, and total P.
                                          Extractable Al had a quantitation limit of 0.021 mg L 1
                                          and a small number of duplicate pairs (n = 11) with
                                          concentrations above  that  limit,  which  would be
                                          expected based on the  low number of acidic lakes
                                          sampled and the correspondingly low extractable Al
                                          concentrations. Most of the field duplicate pairs had
                                          conductance values that were low, because many
                                          dilute lakes were sampled in the WLS-I (Section 5.9).
                                          This factor may account for much of the variability
                                          in the precision of conductance measurements. For
                                          Fe and  total  P, few  field duplicate pairs had mean
                                          concentrations that were above the  quantitation
                                          limit, indicating that the majority of the lakes in the
                                          WLS-I had low concentrations for these variables.
                                          Total dissolved F" had a quantitation limit of 0.016
                                          mg L"1  and a substantial number (n  = 90) of field
                                          duplicate pairs with  concentrations above that limit.
                                          The variability of total dissolved F" seems to be most
                                          apparent within one analytical laboratory (Silverstein
                                          et  al.  1987b). The calculated  precision  for  air-
                                          equilibrated  pH was 0.17  pH  unit compared to an
                                          intralaboratory precision goal of 0.05 pH unit.  This
                                          variability also may  be  related to  the  large number
                                          of  dilute  lakes sampled  in   the WLS-I because
                                          samples from such  lakes are more  sensitive to pH
                                          changes due to the  partial pressure of C02 (Norton
                                          and Henriksen 1983). The precision estimate for DOC
                        34

-------

















CD
CO
CO
a.
>•
*
2
3
U)
CD
CO
—I


2
in
i

« *
(0
CD
Q.
E
W
—
3
4
"to
3
a
Z
2
CD
il
E
g

•D
2
CO
3
_O
o
^
to
2
CO
E
'•^
M
UJ
c
o
2
'5
£
a.
£
u
CO
QQ

cn
c
o
E
^
15
k
CD
O

CO
J)
p







CO
z















in
z
^















z
u.















CO
Z
u_
















a
C CO
o
to
0
ID Q
0_ CO
cc
&v
c
0
^
c fo
^ Q]
a g
c
o
0
Q
c CO
0
o
CD Q
£ to
cc
SR

c
o
'JZ
c g
CD *-
CD C
^ O
C
o
U

Q
c w
0
'to
CD Q
£ CO
or
as

c
o
C 'D
ID t
CD C
5 CD
^ O

o
0

O
C CO
o
!A
'o
CD ._
»- LJ
Q. (O
ff
ss
c
o
Mean
icentratic
o
U


	 i
Q
CO





Variable"
O-J CO CO CN CT> 00 00 O O5 CO
§o to ^~ o o) O) o CN o o o o o CN o ^* o oo r*^ 10 oo co
o en o o <* o •- •- o o o o o o o o o o o «- co o
OO.-OOOOOOOOOOOOOOOOO OOO

•-cococQp~incnocoeneocnoo*-eocMcD CMIO
01 oo . CM m co in r^ C5 05 in CM CM . CM o ^ . *- «- •- i^ ^
COT- inp~ cocMcocn
CM CO CM »-


co in «- co *- CD CM
§*— oco cop^ CM»— cnp^o*— o*~OP^cn incoco
OCOin*-CM*tO
OOinOO'-OOOOOOOOOOOOOO OOO

^»— incMco^cococMOooinop^cM»— «— CM coco
ocococo»— co»— coocop^incM*— incocop*-*. ^ ^- OOP^
CN^I- i- i- ^ r» r^ in OCNOT
«- «- CO «- •-



§co «— p^ ^
co ^ o co ^ ^ o O CM en p^ p^
COCMOOO--OOO cMcoen
OOCO'— OP^*— »— OOOOOO'— OOOP^P*- P^*— O
^ *~ "~



en p^ in oo co
OO'tOO'-CM^^OOOOOOO^OOO O^co
OOCMOO'-OOOOOOOOOOOOOO OOO

^ CM CM co T- co en •* •- P^ CM p> co en p~ ^J- oo •* o in
T— COOCOCDCOOOCMT— ocM^-cncoo^'T- T— in
COi-r- COCO«-T- P»^-~~ ~ *-
£ ""


1^
in CM ^j- oo O «- CM
cnin o*t CM-- p^p~oocor--^tomococo omoo
OO^-CMOcMOOOOOOOOOOOCMof'l' •* ^ «O
CM CO
1 1


CM CO 00 P^ 00 CM
§o COO>-COCOOOOOCMOOOCOCMP^OOOP> encoi-
O^-COt-OOO3OCNOOOOOOOOOOO O^tcM
0 O CM O O «- 0 ^ O O 0 O O O O O O O 0 0 OOO

TJ *>
«— ooooocncnp^p^comoo^'cocMincMP^p^ co^*
co»— in^coi— cnooT~in^cMencM»— ^ co*- ••- ~ oo co
T— in *— CM CM in cocMen *~
«- r- Tj- CM CM


CMCM mm CMOOOO
o-- ^tco oco coocMOocoT-i-ocoin coi-t
• -Q £— —
*-« c CT ~~ .2 >• ? ^ +2
CJCT CD(Dn3"O .t-rCT —
2-0) « " '-e - "DTO®^
^(CS =; •— r- fO — •— Si_D
i§o "° " o S +,'„ 1 « "'5X «™,
_- T- z | i x o - x" i i — ? o















































CO
c
CD
£
CD CO
^. 3 C J2
O CO Q) C
CM ^,^_ jg E CD
II 00 P* r- CD £
^ c CO CO t 1- y,
00 • — ' II II -Q -^ CO 3
UZ-E-E2.2 CUID
r- - -f < S -° E *
-* 1 1 ^ 2 I E
O JO - - C J> Q. I
CDCD-J'-^eDcO °--
citocoigcop, go
3O § CN 2
^5 £ v o> "5 gl
•Xoi^wCD ^CD
(0 — COCOQ— t^.~^
"i-^^wc co^i
EEEEiScD *-ID
ooooJJcj oo
H^ 'f"" -|~ -^~ C (JJ C Q)
"^ ^-^ ^-^ ^— ^ 3 Q_ m J—
co^inco^- r-Q*-™
^^^^'-gco10"
oooo — ^moct/)CD
— i — i — i — i ojo^.t: 3
e - - - E to °0 E 5
"OTJ'OTJ «£^ —
3333CO"*-c5
CO ID CD ID T;, if -t o .2
tOCDCDID**t(]jj: a; c
LU LJ. U_ LJ- < U5 1— Qn
« ja o TJ a »>^

-------
Table 4-4.   Overall Within-Batch Precision Estimates Calculated from Field Duplicates8, Western Lake Survey - Phase I
Variable13
Al, extractable
Al, total
ANC (^eq L~1)
Ca
cr
Conductance (fjS cm"1)
DIC, air equilibrated
DIC, initial
DOC
F~, total dissolved
Fe
K
Mg
Mn
Na
NH4 +
NO3"
P, total
pH, acidity (pH units)
pH, alkalinity (pH units)
pH, air equilibrated
(pH units)
Si02
so4-2

n
189
215
215
215
215
215
215
215
215
213
197
215
215
137
215
83'
205
210
215
215

214
215
215
(x > 0)°
Achieved
Precision (%)e
299.7
33.4
4.5
2.3
15.7
6.5
6.9
7.7
16.7
24.7
112.5
9.6
8.1
139.2
7.2
324.7
61.3
77.4
0.08'
0.08'

0.1 7J
16.1
14.5
Quantitation
Limit
(10SB)
0.021
0.085
24.9
0.30
0.21
6.1
0.71
1.32
2.0
0.016
0.07
0.08
0.03
0.10
0.17
0.13
0.342
0.043
k
k

k
2.07
0.56
(x>10SB)C|d
n
11
11
204
208
65
192
173
103
49
90
16
199
212
1
206
3
8
9
k
k

k
119
124
Achieved
Precision (%)f
44.2
10.2
4.1
2.3
8.6
6.4
6.9
2.7
12.7
26.2
24.6
4.5
8.2
h
7.3
7.9
1.1
37.5
k
k

k
7.0
4.3
Intralaboratory
Precision
Goals
1 0/209
1 0/209
10
5
5
2
10
10
5/1 0s
5
10
5
5
10
5
5
10
1 0/209
0.05 (pH
0.05 (pH

0.05 (pH
5
5


















unit)
unit)

unit)


a Sampling methods (helicopter and ground crew) and laboratories pooled.
bAII units are reported as mg L~1 unless otherwise indicated.
cx is the mean value of a routine and its duplicate.
dSg is the standard deviation of field blank measurements.
eAchieved precision = root-mean-square of the percent relative standard deviation for all pairs with x > 0.
' Achieved precision = root-mean-square of the percent relative standard deviation for all pairs with x > 10 SB
9 Precision goal based on analyte concentration,
hAchieved precision cannot be calculated because x of only one pair of observations exceeds 10SB.
1 Number of observations is smaller due to low NH4+ concentrations in most WLS-I samples and due to instrumental drift (i.e., con-
 centrations of NH4+ < 0).
1 Precision calculated as the root-mean-square of the standard deviation values.
k Quantitation limits are not calculated for pH measurements.
 (12.7%) was based on a large number of sample pairs
 (n = 49)  above the quantitation limit (2.0 mg L~1).
 The high variability may  be due  to the low DOC
 concentrations  present in  the majority  of  lakes
 sampled  in the WLS-I (Section 6.2.7.1). Silverstein
 et al. (1987b) calculated a  precision  estimate of 2.4
 percent using only duplicate pairs having concen-
 trations greater than 5 mg  L"1 (n = 9), while precision
 calculated using pairs with concentrations between
 the  quantitation  limit and 5 mg L"1 (n = 40) was
 14.0 percent.

 Overall within-batch precision estimates calculated
 from field duplicate pairs indicate  how  consistently
 the field samplers  collected  the sample,  how
 consistently the field laboratory personnel processed
 the  sample,  and  how consistently the analytical
 laboratory personnel analyzed the sample. Determin-
 ing within-batch precision on the basis of analyses
 performed at  the five WLS-I field  laboratories
 provides  a measure of performance for this step in
 the sample flow. This precision was calculated using
results from  field duplicate  pairs  and  from field
laboratory  (trailer)  duplicates,  which are field
laboratory splits of routine samples. Table 4-5 shows
the precision estimates for closed system pH, closed
system DIG, turbidity, and true color based on all
sample pairs  with mean  values greater than zero
(x > 0). Quantitation limits could  not be calculated
for  pH and DIG  because  no field  or  trailer  blanks
were analyzed for these variables (Section  2.5.1.2).
The pH and DIG measurements were within desired
precision goals except for the pH of field duplicates
which slightly exceeded criteria.

Precision goals for turbidity and true color were met
when  precision was calculated using mean values
above the quantitation limit (calculated using trailer
blanks) for trailer duplicates. When values below the
quantitation limit were  included in the calculation
(i.e., x > 0), precision goals were  not  met.  Because
most WLS-I lakes sampled had very low values for
turbidity  and true color, poor  precision was not
unexpected.  Based  on  the  results of the field
                        36

-------
 Table 4-5.   Overall Field Laboratory Within-Batch Precision Estimates Calculated from Field Duplicate and Trailer Duplicate Data,
           Western Lake Survey - Phase I

                                                              Precision (Field Duplicates)



Variable
PH
DIC
True Color
Turbidity



Unit
PH
mg L 1
PCU1
NTU'
Intralaboratory
Precision
Goal
(%RSD)a
±0.19
10
±59
10




Achieved Quantitation Limit Achieved
nd Precision (%)e
208 0.129
208 6.8
165 61.3
206 24.4
(10 SB)C nd Precision (%)f
h h h
h h h
32 6 4.7
0.8 37 16.5
Precision (Trailer Duplicates)


Variable
pH
DIC
True Color
Turbidity


Unit
PH
mg L 1
PCU1
NTU'
(x > 0)b
Achieved
nd Precision(%)e
132 0.039
134 3.4
99 25
134 11.8
Quantitation '* '
Limit
(10SB)C nd
h h
h h
15 25
0.35 65
>10SB)b'c
Achieved
Precision (%)f
h
h
3.8
7.8
 a%RSD = percent relative standard deviation.
 bx = Mean values of a routine and. its duplicate.
 °10 SB= 10 times the standard deviation of trailer blanks for trailer duplicates and field blanks for field duplicates
 dn = number of duplicate pairs.
 eAchieved precision = root-mean-square of the percent relative standard deviation for pairs with x > 0 (for pH, root-mean-
 square of the standard deviation).
 'Achieved precision = root-mean-square of the percent relative deviation for pairs with x >10 SB.
 8 Absolute precision goals in terms of applicable units.
 hQuantitation limits cannot be calculated, trailer blanks not analyzed.
 | PCU  = platinum-cobalt units.
 ' NTU  = nephelometric turbidity unit.
laboratory duplicate data, the overall within-batch
precision across the five WLS-I base sites met the
DQOs.
Intralaboratory  within-batch  precision (Table 4-6)
was determined using data from analytical laboratory
duplicates.  The  precision goals  for  laboratory
duplicates were met for every analyte except Mn.
Concentrations of Mn in WLS-I routine samples were
generally low, and the  DQOs do not apply at these
low levels. Therefore,  it is unlikely that a sample
with a  concentration of Mn  high  enough to yield
a  precision  estimate that  would meet the  DQOs
would be selected for duplicate analysis. Based  on
the estimated analytical laboratory precision, the two
analytical   laboratories    showed    excellent
reproducibility.
In conclusion, precision estimates based  on field
audit and  duplicate  sample data  indicate that
samplers, sample handlers, and analysts performed
their tasks consistently.  No systemmatic  precision
problem for any variables  across batches, labora-
tories, or methods was observed.

4.4  Effects  of Holding Time on Sample
Concentration
As part of the calibration study, one of  the  three
samples collected by helicopter crews was selected
randomly to be stored at the field laboratory (in the
dark,  at  4°C)  for as  long as 4 days before it was
processed and preserved (Section 2.6).  The length
of time that the  withheld sample was  kept at the
field laboratory depended on  the amount of time
required  to transport  the corresponding ground
samples from the lake site to  the  field laboratory
(Sections 2.5 and 2.6). This element of the sampling
design  was  intended  to  evaluate the  impact of
delayed  sample  processing  and  preservation that
could result if delivery of the samples collected by
ground crews was  delayed.  The  Forest Service
ground crews, however, were extremely efficient in
providing same-day sample delivery, even from lakes
that were difficult  to  access. Consequently,  the
number of samples for which delivery was delayed
(and thus the number of corresponding  withheld
samples)  was much  smaller than had been  antic-
ipated. In fact, about one-half  of the holding times
for withheld samples was contrived  in order to have
a sufficient number of samples to perform statistical
analyses  on the results of the holding time exper-
iment. The number of withheld samples  assigned
to each holding time  and the analytical laboratories
to which the  samples  were sent are summarized
in Table  4-7. A more detailed discussion  of the
holding time  study  appears  in Silverstem  et  al.
(1987b).
                                                                           37

-------
Table 4-6.   Overall Within-Batch Precision Estimates Calculated from Analytical Laboratory Routine/Duplicate Sample Data,
           Western Lake Survey - Phase I
(x > 0)b
Variable3
Al, extractable
Al, total
ANC (^eq L"1)
Ca
cr
Conductance (^S cm"1)
DIC, air equilibrated
DIC, initial
DOC
F~, total dissolved
Fe
K
Mg
Mn
Na
NH4 +
N03"
P, total
pH, acidity (pH units)
pH, alkalinity (pH units)
pH, air equilibrated
(pH units)
SiO2
so4-2
n
140
148
139
149
149
149
149
149
146
148
144
149
149
120
149
59h
148
137
149
149

149
148
149
Achieved
Precision1*
13.3
3.3
121.6
0.7
2.3
1.1
6.4
4.3
11.1
2.7
56.1
1.4
0.6
105.1
1.1
95.3
8.8
41.7
0.03'
0.03'

0.02'
25.9
1.8
Quantitation
Limit
(10SB)C
0.014
0.029
g
0.04
0.06
3.4
0.33
0.26
0.6
0.008
0.04
0.08
0.02
0.02
0.13
0.07
0.052
0.010
j
j

l
0.28
0.12
(x>10SB)b'c
n
22
64
g
148
145
147
126
131
119
148
63
147
148
45
147
6
118
46
j
i

l
125
143
Achieved
Precision6
4.3
3.3
g
0.7
2.3
1.1
5.8
2.7
4.1
2.7
4.9
1.4
0.6
16.6
1.0
0.9
2.8
7.2
)
i

i
2.9
1.7
Intralaborato
Precision
Goal
1 0/20f
10/20f
10
5
5
2
10
10
5/1 Of
5
10
5
5
10
5
5
10
10/20f
0.05 (pH
0.05 (pH

0.05 (pH
5
5
ry


















unit)
unit)

unit)


aAII units are reported as mg L 1 unless otherwise indicated.
bx = mean concentration of a routine and its duplicate.
c10Se =10 times the standard deviation of calibration/reagent blanks.
dAchieved precision = root-mean-square of percent relative standard deviation for pairs with x > 0.
6 Achieved precision = root-mean-square of percent relative standard deviation for pairs with x >
f Precision goal based on analyte concentration.
9Analytical laboratories were not required to analyze calibration/reagent blanks for ANC.
hn is low due to low NH4+ concentrations (i.e., NH4+ < 0) and instrumental drift.
1 Standard deviation values were calculated for pH measurements.
1 Quantitation limits are not calculated for pH  measurements.
Table 4-7.   Holding Time Samples Processed by Analytical
           Laboratories for the Calibration Study,
           Western Lake Survey - Phase I
Holding Time
(days)
Oa
1
2
3
4
Lab 1
n
19
1
4
1
3
Lab 2
n
6
2
6
3
0
Total
n
25
3
10
4
3
   Total
28
                                   17
                                                45
aZero holding time indicates that samples were processed on
 the date the lake was sampled (i.e., within 16 hours of
 collection).

For  each  variable,  the  difference  between the
concentrations of two of the three samples collected
by the helicopter crew  was calculated.  The two
samples used in this calculation were the withheld
sample  and  the  sample  analyzed  by the  same
laboratory that  analyzed the withheld sample. The
third sample in  each  set was  sent  to the second
analytical laboratory and was not used in the holding
time calculation. For each variable, the variances of
the differences among holding times were equal (i.e.,
they  were  homoscedastic).  The  differences  were
analyzed  as a  function of holding time by  using
standard linear  regression. The number of days that
the  withheld sample  was  stored before  it  was
processed  and  preserved was  the independent
variable in each regression. The dependent variable
was the difference between the concentration of the
withheld sample and the concentration of the  other
sample analyzed at the same analytical laboratory.
The null hypothesis was that the difference between
samples would not change over time. If holding time
had no significant effect, the slope and intercept of
the regression equation would be  zero.

The results of the regression analyses are presented
in Table 4-8. A significant effect of holding time was
demonstrated in only three cases: extractable Al in
samples  analyzed  at Lab  2  (p < 0.032), air-
equilibrated DIG in samples analyzed at Lab 2 (p <
0.010), and air-equilibrated pH in  samples analyzed
                        38

-------
Table 4-8.    Regression Statistics for the Difference Between Routine and Withheld Samples versus Holding Time by Laboratory,
            Western Lake Survey - Phase I
Lab 2
Variable (units)
Al, extractable (/ug L"1)
Al.totaKAigL'1)
ANC (fjeq I/1)
Ca+2 (//eq L"1)
CHA/eqL-1)
Conductance (/jS cm"1)
DIG, air equilibrated (mg L' ')
DIG, initial (mg L"1)
DOC(mgL"1)
F~, total dissolved (//eq L"1)
Fe (/jg I'1)
r
0.243
0.033
0.062
0.054
0.085
0.051
0.330
0.082
0.045
0.003
0.005
0.093
0.010
0.001
0.059
0.013
0.055
0.083
0003
0.005
0.005
0.0001
0.090
P(b)
0.032
0.451
0.304
0.335
0.226
0.350
0.010
0.235
0.385
0.819
0.770
0.205
0.685
0.878
0.318
0.642
0.332
0.231
0.830
0.781
0.779
0.964
0.211
Slope
-0.0008
0.0001
0.169
-0.01 1
-0.007
-0.092
-0.016
0.013
-0.009
0.002
0.0004
-0.0004
-0.004
0.0002
-0.005
0.003
0.012
0.001
-0.010
-0.008
-0.044
0.070
-0.053
Lab
Intercept
1.920
-1.528
-1.180
0.050
0.395
0.010
-0.001
-0.01 1
0.053
-0.421
-2.213
0.026
-0.247
-0.309
0.522
-0.039
0.354
-1.436
0.010
-1 .003
0.024
-0.115
0.187
1
r2O
0.102
0.0004
0.002
0.070
0.046
0.106
0.036
0.024
0.013
0.027
0.013
0.001
0.133
0.041
0.069
0.011
0.004
0.008
0.039
0.024
0.349
0.039
0.113

P(b)
0.103
0.912
0.823
0.182
0.281
0.098
0.340
0.443
0.576
0.417
0.566
0.871
0.062
0.312
0.186
0.610
0.745
0.667
0.326
0.440
0.001
0.321
0.086
 ar = Proportion of total variance explained by the linear model.
 bp = Statistical probability of occurrence using standard linear regression procedure. Values less than 0.05 are considered significant.
at Lab 1 (p < 0.001). For extractable Al, the effect
was  most likely  due  to  one  exceptionally  low
concentration from one sample held for three days.
Because values for all sample pairs were near the
detection limit, no conclusions can be drawn about
the  effect  of holding time on  extractable Al
concentrations.  For air-equilibrated DIG, the effect
was likely due to one  exceptionally  low  value for
this variable analyzed in one sample held for three
days. Removal of the results of the analysis on this
sample, and the fact that the initial DIG regression
showed  no statistical  significance,  indicate  that
holding  time  did not  affect  air-equilibrated  DIG
concentrations.  For air-equilibrated pH, the effect
was due to relatively large differences in this variable
in three samples held for four days. Two of these
sample pairs had differences of 0.1 pH unit. Because
a difference of 0.1 pH  unit is acceptable even for
field duplicate pairs processed at the same time, it
has no practical significance. The third sample pair
showed a difference of 0.5 pH unit; of 45 samples,
this  was the  only one that showed such a large
difference.  This difference probably is  attributable
to random error. Therefore, a practical difference in
holding time  for this variable cannot be concluded
from calibration study data.
The  two chemical variables not included in Table
4-8 are closed-system  pH and closed-system  DIG,
analyzed at the field laboratories.  These analyses
were not performed in the analytical laboratories and
the  number  of  samples  from  calibration  lakes
analyzed in each field laboratory with actual holding
times (i.e., those samples  not analyzed on  the day
they were collected) was insufficient to perform
reliable statistical analyses. Therefore, a regression
analysis was done on each of these variables in the
same manner as for the analytical values except that
all the  calibration lake sample pairs were pooled.
For closed-system  DIG,  the  slope was 0.005, the
intercept was 0.024, the value of r2 was 0.003, and
the value of p was 0.236. For closed-system pH, the
slope was -0.020, the intercept was 0.032, the value
of r2 was 0.032, and the value of p was 0.236. The
results  of this analysis indicate that there  was no
significant effect  of  holding time  on these field
laboratory analyses.
In general, analyses of the calibration study samples
showed no significant effects of sample holding time
on analyte concentration. Possible effects of holding
time, however, were  not  adequately tested  for
samples  with low analyte concentration  (i.e.,
extractable Al and NH4+)  because  reported values
for these samples were near the detection limits.
                                                                          39

-------
4.5  Assessment of the Effect of
Different Field Sampling Protocols

Use of differing protocols involves the risk of bias,
with the  result that data generated  under one
protocol may not be comparable to data generated
under another.  For this reason, the field laboratory
protocols of the survey were  rigidly standardized,
with  holding  times and  handling  procedures
narrowly prescribed. The necessity to alter the field
protocol for wilderness areas surveyed in the WLS-I,
by use of ground crews to collect samples, caused
some concern  about  the effect of the associated
changes in protocol,  particularly sample  handling
and holding times. Differences in the protocols for
ground and helicopter crews are discussed in Section
2.5.1.

A study was designed to assess the differences in
the two protocols  and to  minimize any resultant
effects of the change in protocol. This study required
a random subset of the wilderness sample lakes to
be visited by both ground and  helicopter crews. The
purposes of the study were  to identify (1)  those
variables for which  the values obtained by  the
alternate  ground  crew  method  could  be  used
interchangeably with  values obtained  by  the
standard helicopter crew method, (2) those variables
for which the helicopter values could  be predicted
accurately  from the  ground values, and (3)  those
variables for which the  ground  crew  method
produced results that  were unacceptable for use in
data analysis. The second  aspect of the approach
involved calibrating the ground crew  results with
those  from the helicopter crews.  Used in this
context,calibration  is simply adjustment of  the
ground crew values using a linear regression model
to make them comparable to  data collected by the
helicopter crews.

Of the 50 lakes selected for this calibration study,
45 were actually sampled in duplicate by both ground
crews and  helicopter crews. Visits were scheduled
for  the same day, but in order  to ensure data
comparability, ground crews were not told which of
the lakes also  were to be sampled by helicopter
crews. The identity of the calibration  lakes inadvert-
ently was  made known to the field crews in  one
subregion before sampling began. The sample for
this subregion  was redrawn, and,  in  order not to
indicate the newly selected lakes, duplicates were
not reassigned for the ground crews. For this reason,
several sample  lakes do not have duplicate ground
sample values. Additionally, some  data  were
removed from  the analysis  as  a  result  of data
validation (Section  3.3).

In addition to the use of two methods of field sample
collection, the survey also  employed two analytical
laboratories. The duplicate samples collected for the
calibration study provided evidence of the repeat-
ability  of the  analytical  measurements,  which is
expressed as measurement error.  One sample of
each duplicate  pair was sent  to each  of the two
analytical  laboratories  contracted  for the survey.
Additionally, information regarding laboratory bias
was provided by audit samples (Section 4.3) and field
laboratory splits.

4.5.1  Calibration of the Ground Sample Data

The objective of the calibration study was to identify
those variables for which (1) the two field  sampling
protocols generate interchangeable results, (2) those
for which the helicopter values are different, but are
predictable from the ground crew values, and  (3)
those for which such prediction involves unaccepta-
ble error, so that the  ground  crew data must  be
considered  unusable  in generating  population
estimates. Details of this analysis are  provided in
the Appendix.

Analysis began by plotting the helicopter value (y)
againstthegroundvalue(x; Figures4-1 to4-3). Lakes
with duplicates for each variable are represented by
a box, with the four possible pairings of ground data
with helicopter data  forming each  corner. Missing
values result in missing sides of some boxes. When
the two protocols yield identical  values for  the
variables,  the results appear as points on the one-
to-one, diagonal line.

Based  on  inspection  of these plots, a linear model
of the  form  y = bx  + e was used  to compare the
sampling  methods.  The model  results  (Table 4-9)
indicated  that  the  differences between samples
collected by helicopter  and by ground crews were
small,  a conclusion that was also supported by the
plots of the paired data.

The slopes of the regression between helicopter and
ground samples for the  primary variables were close
to one, although statistically significant differences
were observed for S04 2, pH, and  DOC.  However,
calibration of the ground data  using the helicopter
data provided little improvement for SO«~2 and DOC.
This finding also was  evident when the relative
variances  (total  variance/population variance)  of
helicopter data were compared to these for  the
calibrated data (Table 4-9).

For extractable Al and pH, differences were apparent
between the samples  collected by ground and  by
helicopter crews.  However, calibration  of  these
variables,  although  statistically defensible, is  not
warranted considering  the  nature of the  data. For
pH, maximum bias caused by  use  of ground crew
samples is decreased only 0.001  unit (from 0.007
to 0.006) by use of the calibration model. This change
in hydrogen ion concentration would be small in
                      40

-------
Figure 4-1.    Results of ANC (fjeq L"1) measurements and pH determination on air-equilibrated samples on duplicate samples
              collected by helicopter and ground crews from  45 lakes for the calibration study during the Western Lake
              Survey - Phase I. The dimensions of the boxes indicate the magnitude of sampling and measurement error;
              sampling bias is evident when the boxes consistently deviate from the 1:1 (diagonal) line.
                              650
                          -j
                           o-
                           .
                           
-------
Figure 4-2.    Results of sulfate (fieq £"V and calcium (/jeq L'^) analyses on samples collected by helicopter and ground crews
              from 45 lakes for the calibration study during the Western Lake Survey - Phase I. The dimensions of the boxes
              indicate the magnitude of sampling and measurement error; sampling bias is evident when the boxes consistently
              deviate from the 1:1 (diagonal) line.
                             140
0)
!

                             70
                                                                                       140
                                                   Ground Sample (fjeq L ^j
                             600
                         -S

                         %   300
                         &

                                       Ca
                                                            300

                                                   Ground Sample faeq i"
                                                                                       600
                           42

-------
Figure 4-3.    Results of extractable aluminum ffjg £~V and DOC (mg L V measurements on duplicate samples collected by
              helicopter and ground crews from 45 lakes for the calibration study during the Western Lake Survey - Phase I.
              The dimensions of the boxes indicate the magnitude of sampling and measurement error; sampling bias is evident
              when the boxes consistently deviate from the 1:1 (diagonal) line.
                           I
                           0)

                           I
                           10
                            *
                            01
                            I
                           .o
                           1
20
                                10 '
                                              10         20         30

                                                       Ground Sample (/jg L~
                               6.5
                           I
                           5
                              325
                                       DOC
                                                            3.25

                                                      Ground Sample (mg L'
                                                         6.5
                                                                                     43

-------
Table 4-9.    Results of Regression of Values for Primary Variables Measured on Samples Collected by Helicopter Crews
           (Dependent) Against Samples Collected by Ground Crews for the Calibration Study Lakes, Western Lake Survey -
           Phase I"

                                                                   Relative Variance15
Primary
Variable
ANC
pH (closed system)
Ca + 2
so4-2
DOC
Al, extractable
Slope
0.99961
0.99313d
1.00261
1.01314d
0.97454d
1.05926
r2
0.9991
0.9997
0.9982
0.9979
0.9912
0.5889
Helicopter
Protocol
1.00108
1.00976
1.00156
1.00290
1.02431
1.09462
Calibrated
Ground
Protocol
1.00123
1.06918
1.00111
1 .00390
1 .02909
1.46587
Direct Use of
Ground Protocol0
e
1.07856
e
1.00395
1.03010
e
a Regressions, based on a sample size of 45, were forced through zero and thus no intercept term is shown.
bRelative variance is defined as total variance/population variance where total variance is the sum of population variance, measure-
 ment error variance, and prediction variance.
0Shown only for those variables where the relative variance resulting from calibration is less than direct use of ground protocol.
d Significantly different from 1.00 (at p<0.05).
eNot possible to compute.
acidic lakes, and would be unnoticeable at the high
pH values  encountered in  the western lakes.
Although the slope for extractable  Al  showed no
significant bias between the two protocols, extract-
able Al  based on  results from the ground crew
samples were much more variable than those based
on helicopter crew samples, with relative variances
of 1.46587 and 1.09462, respectively. This increased
variance between the two protocols would normally
indicate the need for calibration. However, most of
the extractable Al concentrations were below the
detection limit, and those concentrations above the
detection limit were far below any level of concern
for biological effects associated with aluminum
toxicity.

Consequently, samples collected by ground  crews
were judged  acceptable  for  all variables and no
calibration or replacement was conducted. Based on
the result of this analysis, the results generated from
the ground crew sampling protocol were combined
with those generated from the helicopter sampling
protocol. All population estimates presented in this
report are based on the combined data.

4.5.2   Relative Bias Between Analytical
Laboratories

The constraints of processing  a large number of
samples within the prescribed holding times for the
measured chemical  variables  necessitated the use
of two   analytical  laboratories  in  the WLS-I. To
evaluate systematic differences in performance
between laboratories, three groups of samples were
collected: (1)  synthetic  and  natural  audits, (2)
helicopter and ground duplicate samples from the
calibration study lakes, and (3) field laboratory splits
of all samples analyzed  by  EPA's  Environmental
Research Laboratory at Corvallis. These three types
of samples are referred to as audits, duplicates, and
splits, respectively. Methods of sample collection and
processing are described in Sections 2.5 and  2.6.
Analysis of each type of these QA samples provided
the opportunity to examine relative bias between the
two analytical laboratories (Lab 1 and Lab 2).

Results  of  analyses  on  audits were analyzed
differently  from those  derived  from analyses on
duplicates and splits because the means of six audit
lots constitute the "sample." Estimates of variance
of each datum provide the basis for weighting each
of these  samples,  using a maximum  likelihood
approach (Permutt and  Moezzi 1986).  Data for the
duplicates  and splits were analyzed as individual
records because these data are not amenable to
maximum likelihood analysis. The duplicate and  split
samples also  differ from  one another  because the
duplicates constitute pairs of samples from the same
lakes sent to both laboratories; the split samples  also
were sent to a third analytical laboratory so that EMSI
and  Versar do  not  appear together in  the same
sample pair.

Despite the differences  among the three types of
QA samples,  relative bias between the analytical
laboratories was assessed using a  weighted least
squares model, with some modification. First, the
conventional regression analysis (shown as Model
2) can  be  modified  as  shown  in  Models 0 and 1
where:

    Model 0:  y = a + x + e  (b=1)
    Model 1:  y = bx + e  (a = 0)
    Model 2:  y = a + bx + e

These three  models were used to describe  the
differences between laboratories where Model 1  was
the model of first choice if more than one model
satisfactorily described the  relationship.  The
parameter estimates for the audit samples using the
least squares approach were very  similar to those
                       44

-------
Table 4-10.    Estimates of Bias between Analytical Laboratories Based on Synthetic and Natural Audit Samples (Audits),
             Paired  Duplicate Samples from the  Calibration Study (Duplicates),  and  Split  Samples (Splits) from the
             Environmental Research Laboratory-Corvallis. The Model Results are Shown Using Lab 2 as the Dependent
             Variable and Lab 1 as the Independent Variable
Variable
Primary
ANC
Ca+2
so4-2
DOC
Al, extractable
Secondary
Mg + 2
Na +
K +
NH/
N03-
cr
F-
DIC, initial
DIG, air equil.
pH, air equil.
pH, acidity
pH, alkalinity
Al, total
Fe
Mn
P, total
Si02
Conductance
Unit

#eq L~1
Aieq L~1
/ueq L'1
mgL"1
/ugL'1

/ueq L-'
/ueq I"1
,ueq L"1
yueqL'1
yueq L"1
/ueq L"1
Veq L'1
mgL-1
mgL"1
pH units
pH units
pH units
//gL-1
MJL'1
«JL-1
A/gL'1
mgL-1
,uS cm'1

Audits'1

NS
(0)
(0)
0.17
-1.9

(0)
NS
0.67
0.55
(0)
NS
(0)
NS
NS
(0)
(0)
NS
5.9
-8.3
NS
-1.1
(0)
(0)
Intercept3
Duplicates0

-5.9
(0)
(0)
0.21
NS

-0.46
NS
NS
-0.414
0.36
NS
(0)
(0)
-0.11
1.34
-0.1317
-0.90
NS
-3.54
3.74
2.39
NS
NS

Splits0

NA
(0)
(0)
NA
NA

NS
NS
NS
NA
NA
NA
NA
NA
NA
NA
NA
NA
7.12
NS
1.72
NA
NS
NA

Audits'

NS
1.054
0.945
0.948
0.68

1.016
NS
0.925
0.16
0.959
NS
1.048
NS
NS
0.9952
1.0123
NS
0
(1.0)
1.041
NS
1 .0984
0.989
Slopeb
Duplicates0

1.055
1.073
0.970
0.853
1.272

1.056
NS
NS
(1.0)
0.788
NS
0.913
1.050
1.085
0.807
(1.0)
(1.0)
0.721
0.840
0.133
0.589
NS
NS

Splits0

NA
1.077
NS
NA
NA

NS
NS
0.949
NA
NA
NS
NA
NA
NA
NA
NA
NA
0.904
NS
0.850
NA
NS
NA
a Intercepts shown as (0) indicate that Model 1 was used in the analysis.
bSlopes shown as (1.0) indicate that Model 0 was used in the analysis.
CNA = not analyzed, NS = not significant.
derived from  the  maximum likelihood method  in
Permutt and Moezzi (1986).

The  results of the analysis of analytical laboratory
bias using the  audits, duplicates, and splits are
shown in Table 4-10. Cases evaluated using Model
0 are indicated with slopes (b) set to 1.0. Those cases
analyzed with Model  1 are designated  with the
intercept (a)  set to  0.  Those variables with both
intercept and slope significantly different from 0 and
1, respectively were fitted with Model 2. All analytical
variables, except sodium and chloride, exhibited  at
least one significant regression parameter for the
QA comparisons between laboratories. Estimates  of
bias in the  intercept term are generally inconsistent
among the three types of QA samples. Also, the
magnitude  of the  estimates of intercept bias are
relatively small. Several exceptions are DOC, air-
equilibrated DIC, pH  (acidity) and Fe. In all  of these
cases,  the magnitude  of  the bias term  for the
intercept,  although statistically  significant,  is
unimportant with respect to potential effects on data
interpretation. In contrast,  the  estimate of bias for
air-equilibrated pH of 1.34 units based on  analysis
of the duplicates is large. However, the analysis  of
the audit samples indicate no  problems with bias
for this variable and removing two duplicate samples
with aberrantly high deviation (Figure 4-4) produces
results consistent with the analysis of the audits.
Estimates of slope show consistent bias for Ca+2,
S04~2, DOC, N03", DIC and air-equilibrated pH.

Differences in ANC are also  evident between the
laboratories (Figure 4-4), but at concentrations less
than 100 /ueq L"1, these differences were very small.
At higher values of ANC, the differences observed
between laboratories were generally smaller than
the difference in split samples measured within a
single laboratory.

The maximum difference observed between the two
laboratories for  air-equilibrated pH  measurements
was approximately 0.4 pH  unit. Possible sources of
error in this pH measurement include: (1) incomplete
equilibration with 300 ppm COa in air, (2) electrode
errors, or (3) differences in measurement procedures
between the laboratories. Many of the lakes sampled
in the calibration study were very dilute, and the
pH is sensitive to changes in partial pressure of COz.
Small  differences in the  COa  gas  used  at each
laboratory in equilibration, the equilibration  times,
or measurement procedure, coupled with electrode
                                                                           45

-------
Figure 4-4.    Results of ANC f/jeq L'^) and pH determinations on air-equilibrated samples performed by Lab  1 and Lab 2.
              Analyses were conducted on duplicate samples collected by helicopter and ground crews from 45 lakes for
              the calibration study during the Western Lake Survey - Phase I. The dimensions of the boxes indicate the magnitude
              of sampling and measurement error; laboratory bias is evident when the boxes consistently deviate from the
              1:1 (diagonal) line.
                               700
                              560
                            •8
                              420
                              250
                               140
                                              740
                                                         2SO
                                                                    420
                                                                                560
                                        700
                               5.5
                                              6.1
     6.7         7.3
Air-Equilibrated pH - Lab 2
                                                                                 7.9
8.5
                           46

-------
 errors, could produce the observed differences in the
 air-equilibrated pH values. The air-equilibrated pH
 measurement was not used in deriving population
 estimates, and thus no corrective action was taken.

 For  Ca+2, the three  parameter estimates of slope
 show that Lab 1 measurements are between 4 and
 8 percent greater than those from Lab 2 (Figure 4-5).
 If Model  1 is used for Ca+2 on the audit samples,
 the  estimate for bias in the slope approaches  that
 observed for the duplicates and splits. Differences
 in calcium can be attributed to different analytical
 protocols that were allowed under the EPA contract;
 Lab  2 used atomic absorption spectroscopy and Lab
 1  used  inductively coupled plasma emission
 spectroscopy.

 Differences between the results of analyses for the
 following variables cannot be attributed to differen-
 ces  in  instrumentation.   For  S04 2, the Lab 1
 measurements appear to be from 2 to 5.5 percent
 less than those from Lab 2. Nitrate analyzed by Lab
 1  also appears to be somewhat lower than Lab 2
 measurements (Table 4-10). The  analytical labora-
 tory measurements of DIG show that  Lab 1 values
 are greater than those from Lab 2 by approximately
 5to  13 percent. Extractable aluminum results (Figure
 4-6) illustrate the high degree of variability typically
 observed for variables with most of the values near
 the  detection limit.  Dissolved  organic  carbon,  in
 contrast,  shows a much higher degree of precision
 that allows  differences between laboratories to be
 detected (Figure 4-6).

 Given the  presence of  statistically significant
 differences between analytical laboratories, it would
 be desirable  to adjust  for  measured  differences.
 However, deriving an appropriate adjustment for
 analytical bias involves a high degree of uncertainty.
 The  results from the three types of QA samples show
 considerable  lack of  agreement  in the estimated
 magnitude of the relative bias. An additional problem
 in adjusting for analytical  bias is that at least nine
 variables show a significant lack of fit with the linear
 models, as illustrated in the analysis of the audits
 (Permutt  and  Moezzi 1986).  As shown  in  the
 calibration  study   (Appendix),  correction  for
 differences in protocol also introduces an additional
 source of error.

The  audit,  duplicate, and split sample results
presented here to evaluate relative bias between
laboratories  cannot be used to judge how well  the
survey results met the DQOs for absolute accuracy.
Determination of absolute accuracy requires that the
results be compared to certified standards of known
concentrations, e.g.,  from  the  National Bureau of
Standards. Because the use of certified standards
was  beyond the scope of the QA program for  the
WLS-I, such  a  comparison is not possible. Never-
theless, it is possible to show that the relative bias
between laboratories for most variables is either not
statistically significant, or  not  meaningful in the
context of the survey objectives. However, because
it is difficult  to establish absolute accuracy for an
analyte, correcting for the observed  relative bias in
this survey  requires  more  information  than is
currently  available.  Consequently,  the  data pre-
sented in this  report have not  been adjusted for
relative bias.

4.6   Assessment of Data Quality

Data quality is a measure or description of the types
and magnitude  of error associated with a data set.
The quality of a data set can be expressed  in terms
of five characteristics: precision, accuracy, represen-
tativeness, completeness, and comparability.

The QA data  results presented  in Section 4 allow
the data user to evaluate how well the characteristics
of  each sample  are  represented by the  data
associated with the lake sample collected in the
WLS-I. To make this determination, the QA data are
compared to the  data quality  objectives (DQO)
established before the survey was initiated (Section
2.1.1).

Of the issues that  can be addressed by comparing
WLS-I data to the initial DQOs, precision may provide
the most extensive information.  Precision  was
calculated by  using two types of field audit samples
and by using duplicate samples (Section 3.5.2.2). The
use of the two types of audit samples is an effective
means of determining analytical precision. Water
samples collected from lakes (natural audits) can be
used to estimate how  well the variables within a
natural matrix are repeatedly analyzed by the same
laboratory  and  by different laboratories. Water
samples prepared from a matrix of known analytes
and deionized water by a reference laboratory can
be  used to estimate the precision with  which the
analytical laboratory measures the analytes known
to be present in the sample. The overall among-batch
precision estimated on the basis of more than 200
of these field  audits was excellent both within and
between the WLS-I analytical laboratories.

Overall within-batch  precision estimated from
duplicate samples is another means of determining
confidence  in  sample values. For  the WLS-I,
duplicate sample values were compared from the
time the  lake  was sampled  until  the  analyte
concentrations  were reported  by  the  analytical
laboratory. For the  WLS-I duplicate samples, when
the quantitation limit  is used  as  a criterion  to
eliminate pairs that are of low concentration, within-
batch precision  can be shown to be well within the
acceptable limits established for sampling, process-
                                                                        47

-------
Figure 4-5.    Results of suit ate (fjeq L V and calcium (fjeq I"1 measurements by Lab 1 and Lab 2. Analyses were conducted
              on duplicate samples collected by helicopter and ground crews from 45 lakes for the calibration study during
              the Western Lake Survey -  Phase I. The  dimensions of the  boxes indicate the magnitude of sampling and
              measurement error; laboratory bias is evident when the boxes consistently deviate from the 1:1 {diagonal) line.
                           •8
                           1
                           
-------
Figure 4-6.    Results of extractab/e aluminum (fjg i"V and DOC (mg L 1^ measurements performed by Lab 1 and Lab 2. Analyses
              were conducted on duplicate samples collected by helicopter and ground crews from 45 lakes for the calibration
              study during the Western Lake Survey - Phase I.  The dimensions of the boxes indicate the magnitude of sampling
              and measurement error; laboratory bias is evident when the boxes consistently deviate from the 1:1  (diagonal)
              line.
                                50
                                40
                                              /O         20          30

                                                      AI-Ext. (fjg L-') - Lab 2
                                                      40
                                                                 50
                               6.5
                          •Q
                          10
                          -J
-J
I
O
Q
                             325 -•
                               0.0
                                                             3.25
                                                      DOC (mg L'') - Lab 2
                                                                                           6.5
                                                                                       49

-------
ing, and analytical error. For field duplicates, most
analytes showed precision within or near the desired
limits. Those analytes with precision slightly outside
the acceptable limits should not have a large effect
on overall population estimates, but their impact on
interpretation must be considered by the data user.
Values for all  other analytes, based on  trailer
duplicates and laboratory duplicates,  indicate
consistency in all  steps from sample  collection at
the lake  site to preparation of  the  verified  and
validated data sets.

Accuracy  is sometimes defined as the combined
components of precision and bias (Kirchmer 1983);
but, for the purpose of  assessing data quality it is
used here synonymously with the term "bias." Bias
can be introduced through systematic contamination
of the samples, through  differences in  sample
collection, or through differences in treatment and
analysis of the samples.

Except for random error, the samples analyzed during
the WLS-I  were minimally influenced by contam-
ination.  Most  of the  field blank values  were
sufficiently near the detection limits to conclude that
no  systematic  contamination  in field sampling,
preparation of sample aliquots, sample processing
and transfer, or sample analysis occurred. Analytical
results from trailer blanks, laboratory calibration and
reagent blanks  also  indicate  that there was no
systematic contamination in the field  laboratory or
in the analytical laboratory. As a  result, population
estimates are not influenced by contamination of the
samples.

However, bias was introduced in the  data through
differences in helicopter and ground access to the
lakes  and through the use  of two analytical
laboratories.  Bias resulting from differing  field
sampling  protocols was found  to  be statistically
insignificant  for most  variables, and  marginally
significant for the remaining variables. Bias intro-
duced  by the laboratories  was observed  for most
variables, but absence of an unambiguous standard
prohibited calibrating the data to minimize analytical
laboratory bias. However, the  accuracy  of the
measurements of cations and anions can be checked
by comparing the calculated equivalent conductance
of the sums of ions with  the measured  conductance.

The measured conductance values between the two
laboratories agreed very well. The  only significant
difference observed  among the three types of QA
samples  (Table  4-10) was a slope of 0.989 for the
audit samples.  However,  using a least-squares
regression  through the origin, the calculated
conductance values from Lab 1 on  those from Lab
2 yields a significant slope of 1.049.  This slope is
almost identical to that obtained when regressing
the sum of cations (1.055) or  the  sum  of anions
(1.048) for  Lab 1 on Lab 2. This  indicates that  Lab 1
is consistently overestimating anions and cations by
approximately 5 percent (within the range of samples
in the calibration study).

A sensitivity analysis was conducted  to determine
the  impact  of adjusting  calcium for  bias, the
parameter  showing  the greatest  degree of bias.
Population  estimates for the percent of lakes in each
subregion with  calcium < 50 /ueq L"1  changed less
than one percent when a bias correction was applied
to  the data by subtracting 10 percent from Lab 1
values.  Thus, the  impact of  analytical bias on  an
objective of the survey, namely characterizing the
chemistry of lakes, appears minimal.

Other analyses, such as those presented in Section
6, are potentially more sensitive  to misinterpretation
caused  by  analytical  bias. Again, a "worst-case"
sensitivity analysis was conducted by subtracting  10
percent from the calcium values from Lab 1, the
laboratory  that  analyzed  the  samples  from the
Central  (4D) and Southern Rockies (4E). A  linear
regression  of ANC on  the  sum  of [Ca+2 +  Mg+2]
showed  the  parameter  estimates  for these two
subregions to exhibit  minor changes  in intercept
between the adjusted and unadjusted data, but more
appreciable differences in the slopes  (Table 4-11).
Although these differences  in slope  represent a
measurable change in the relationship between ANC
and  [Ca+2 + Mg ],  they have little impact on the
interpretation of these particular analyses. Users of
the  data should  be  cognizant  of the  possible
limitations  in the  data  attributed to differences
between laboratories.
Table 4-11.    Comparison of  Model  Results of ANC  on
             [Ca+2 + Mg+2] for the Central and  Southern
             Rockies Using Data  with Calcium Adjusted
             and Unadjusted for Possible Analytical Bias9,
             Western Lake Survey  - Phase I
Subregion
Central Rockies (4D)
bias adjusted
unadjusted
Southern Rockies (4E)
bias adjusted
unadjusted
Intercept
-8.9
-9.9
-16.0
-11.4
Slope
0.979
0.935
1.085
1.007
 aCa + 2 was adjusted by reducing the reported values from Lab 1
 by 10 percent. A linear regression model weighted for the
 inverse sampling probabilities was used
Representativeness is the degree to which sample
data accurately and precisely reflect the character-
istics of a  population,  i.e.,  how  well  the WLS-I
samples characterize the lakes in the region (Section
2.1.1).  Representativeness  has several  aspects
including:
                       50

-------
 1.  how well a single water sample collected at
     1.5  m over the deep portion  characterizes  a
     lake (spatial representativeness);

 2.  the  degree to which a  sample  on  one day
     characterizes the nature of that lake (temporal
     representativeness); and

 3.  the  degree to  which the sample  of  lakes
     represents the population of  lakes (sampling
     representativeness).

The assumption of spatial representativeness is that
the water mass in a lake is chemically homogeneous,
a condition most likely to exist when lakes are non-
stratified, or mixed. Nearly all lakes  were confirmed
to be nonstratified through  the measurement  of
temperature profiles.

Sampling representativeness  cannot be known
exactly without conducting a complete census of the
target population. However, the sampling  scheme
was  designed to maximize the representativeness.
A  systematic random sample drawn within  strata
ensured good geographical coverage without bias.

Completeness is  a necessary  part of evaluating
population  estimates because valid statistical
evaluations cannot be performed unless a sufficient
number of samples are  collected and analyzed. A
goal  of the survey was to sample approximately 50
lakes per stratum to allow population  estimates to
be computed with a high degree of precision. Each
of the five subregions had three strata  for a desired
total target of  approximately 750 lakes  for the
probability  sample.  However,  only  720 probability
sample lakes were sampled; five strata where the
achieved sample size was less than 90 percent of
the sampling goal  were  4A3, 4C3, 4D1, 4D3, and
4E3  (Table 5-5). Only in Subregion  4D were two of
the three strata inadequately  sampled; this lack of
completeness caused concern  regarding the  popu-
lation estimates in Subregion 4D.
the calibration study and other QA samples failed
to identify  any serious problems in comparability
among field crews or field laboratories. The subject
of comparability between laboratories was discussed
earlier in Section 4.6.

The evaluation  of the quality of the WLS-I data is
dependent  upon  its  intended use; therefore,  the
quality was assessed from the perspective of  the
objectives stated in Section 1. For  the purposes of
preparing population estimates and characterizing
the nature  of these lakes, the quality of the data
is judged to be acceptable. Other users of the data
may wish  to  reevaluate the  quality  of the data
consistent with its intended use and their objectives.
Comparability refers to the similarity of data from
different  sources included in a single data set.
Different sources used to collect and generate the
data for the survey include 68 field  crews used to
collect the samples and make in situ measurements,
five  field  trailers used to make field  laboratory
measurements, and two analytical laboratories used
to analyze the samples (not including split samples
sent to ERL-Corvallis).  Because of the large number
of field crews  used to collect the samples, it was
not  practical to provide a sufficient  number of QA
samples  to quantify  differences among crews.
However,  the common training and  equipment
provided to all crew members served to minimize
this  source of variability. Available  evidence from
                                                                        51

-------
                                           Section 5
                              Results of Population Estimates
5.1   Data Presentation and
Considerations

5.1.1  Presentation

In the interest of condensing the results, most tables
show data for regions and subregions, omitting the
results of the  alkalinity map class strata (Section
2.2.1).  Summaries of subregional population
estimates are provided for California (4A), the Pacific
Northwest  (4B),  the Northern Rockies  (4C), the
Central Rockies (4D), and the Southern Rockies (4E).

Only  the primary variables,  acid  neutralizing
capacity, pH, sulfate, calcium, extractable aluminum,
and dissolved  organic carbon,  are  summarized  by
extensive tables in this section. Of  the variables
measured,  these  were selected because  of their
direct relevance to issues of acidic deposition effects.
Nitrate may also be important, but it  was present
in low concentrations. Lake population statistics are
also provided for estimates of the physical attributes
of the lakes  sampled, as  well as for secondary
variables  of interest. Summary statistics  for  all
measured variables are presented  in Volume II  of
this report.

The descriptions of each population were summar-
ized on  a single page of output for each variable.
The cumulative distributions, F(x) and  G(x), supple-
mented by descriptive statistics, serve as the primary
data outputs and were used to generate the tabular
information presented in this report. An example of
these outputs  is shown in Figure  5-1  and Volume
II includes these outputs for  other variables and
subpopulations.

Each  page of data output (Figure 5-1) is identified
at the top by the name of the variable (chemical or
physical  parameter) and by the definition of the
subset of lakes being examined. A subset can  be
defined  on the basis of  any physical or  chemical
variable or design  stratification  factor  (region,
subregion or alkalinity  map class), either singly or
in combination.

The basic population estimates, N  and A (denoted
only as  N and A on the output), and the standard
errors of these estimates are presented (Figure 5-1)
in the heading of the output. N is  defined as the
estimated total number of lakes in  the population
being described, and A  is the total amount of lake
area (ha) for all lakes in that population. The indicated
sample size (n***. Section 2.3.2) is the number of
sampled  lakes  in the target population  being
described. The identity of reference values (Xc1, Xca,
XC3> chosen for comparative analyses of that variable
completes  the heading  material on each page of
output.

The distribution curves,  F(x) and G(x), are similar in
interpretation, representing  estimated frequency
and areal distributions,  respectively. For any value
x of variable X, F(x) is interpreted as the estimated
proportion  of lakes having^ a value of X < x. F(x) is
calculated  as the ratio of Nx, the estimated number
of lakes having a  value of X < x, to N. The upper
95 percent ^confidence  limit on NX,  also scaled by
dividing by N, is shown as a dashed line (Figure 5-1).
Although  only  a one-sided upper 95  percent
confidence limit  is shown, a  two-sided 90 percent
limit can be made by using the existing limit and
one equidistant  below  the curve. This procedure
would also provide a lower 95 percent one-sided
limit.

The function G(x) is the estimated proportion of lake
area in the population having a value of X < x, for
any variable X. This cumulative areal curve and its
confidence curve are determined analogously to the
F(x) curves.

An example of interpreting these  distributions is
derived from Figure 5-1. At the value of ANC = 200
/ueq L~1, F(200) = 0.67 (calculated and shown on the
figure as pca = 0.666); the upper confidence limit
is  read from the dashed  line as 0.70. These values
are multiplied by the estimated total number of lakes
in  the population,  10393, to yield 6859 (calculated
and shown on the figure  as Nc3 - 6926) as the
estimated number of lakes in the population having
ANC < 200 //eq L~1, and an upper confidence limit
for that number, 7275 (calculated and shown on the
figure as NCU3 =  7276).  The calculated  values and
those derived from the figure differ due to the
imprecision of reading  the proportions (0.66 and
0.70) from the curves.  Reading values from the
                                               52

-------
Figure 5-1.    F(x) and Gfx) distributions f •
             •) and the 95 percent upper confidence limits (
                                                     ) on these distributions
              for ANC (fjeq L  ) for the target population of lakes f >  1 ha and < 2000 ha) sampled in Region 4 (West).
              Western Lake Survey - Phase I. The population statistics, quintiles (Q, - Q« ) and medians, and three reference
              values (Xci. Xca,  and XC3 ) for ANC are also shown. The  proportions (pc or 
-------
curves is suitable only for obtaining rough estimates,
due to the potential error that can be introduced.
Precise computation requires the use of the  data
base and the appropriate algorithms.

Plots of F(x) and G(x) can be  used only to estimate
confidence limits on the number or area  of lakes
below a reference value of X. When the reference
value for variable X might best be discussed in terms
of the population greater than  x,  the plots  are
presented as  complementary distributions, 1-F(x)
and 1 -G(x), which can be computed for any variable
of interest. Examples of these distributions are found
in Section 5.4.3 for sulfate and in Volume II.

Associated with  each distribution  shown  on  the
output are the following descriptive statistics:

• Mean  and  standard deviation,  appropriately
  weighted to reflect the estimated distribution, F(x).
  There are no comparable statistics for G(x).

• Min and  Max are the minimum and maximum
  values  of variable X observed in samples  for the
  population or subpopulation being examined.

• The median is the estimated value of variable X
  such that half the number of  lakes, or half the
  lake area, in the  population is at or below this
  value.

• The quintiles  (Qi, Q2, Q3, Q4, Qs) are the values
  of X that partition the distribution  into five equal
  parts (e.g., the 4th quintile is the 80th percentile;
  Q5 is the maximum value). These values are useful
  when  comparing  differences  in  F(x) or  G(x)
  distributions  among regions, subregions,  or
  subpopulations. The  interquintile  difference (Q Xc).

                       54
5.1.2  Design Considerations

5.1.2.1   Using Weights—The design of the WLS-I
requires that the results be presented as population
and/or subpopulation estimates whenever conclu-
sions on combined strata  are to be drawn (Section
2.3). Expansion factors or weights (W) must be used
when making combined strata estimates of attributes
for the populations of lakes (Linthurst et al. 1986).
These weights  are  defined,  and  the  estimating
equations are given, in Section 2.3.

Using Strata 4A1 and 4A2  illustrates the require-
ment that all unweighted estimates be made within
strata and that means or  other statistics involving
more than  one stratum  be calculated with  the
appropriate stratum weights (Table 5-1). The correct
way to estimate the total number  of lakes in two
strata below a reference value (in this example ANC
< 50 //eq L"1) 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
reference value for each stratum (nc/n***: 23/54
= 0.426  and 12/53 = 0.226). 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
estimated number of lakes in the stratum population
below the reference value. Adding  the Nc for each
stratum (735.5 + 101.1) yields the combined stratum
Nc (836.6). The same answer can  be obtained by
multiplying nc by W for each stratum and summing
the results.

The most accurate estimate for the overall proportion
of lakes in  the designated population below  the
reference value, therefore, is 836.6/2173  = 0.384
(Table 5-1).  If the overall proportion of lakes below
Table 5-1.   Use of Weights in Combined Strata Estimation,
          Western Lake Survey - Phase I

                                ANC < 50 Meq L"1
Stratum
4A1
4A2
Combined
N
1727
446
2173
n***
54
53
107
W
31.978
8.422

nc
23
12
35
Pc
0.426
0.226
0.384
Nc
735.5
101.1
836.6
   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 ANC
       < 50 ^eq L~1, the reference value.
  pc = estimated proportion of lakes in the sample (for a
       stratum) or population (for combined strata) which has
       ANC < 50 /jeq L~1 (nc/n***).
  Nc = estimated number of lakes in the population which has
       ANC < 50 f^eq L~1, the reference value.

-------
the reference value were  computed  as 35/107 =
0.327 (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 2173 lakes in Strata
4A1 and 4A2. Using the correct value of 0.384 as
PC, the  estimated number of lakes with ANC < 50
/jeq LT1  would be 837. Using the  incorrect pc value
of  0.327  (based on  the  combined nc/n***),  the
estimated number of lakes with ANC < 50 peq L"1
would  be  711. Therefore, the  number of lakes
estimated to  have ANC < 50 /ueq L"1 in both strata
would be underestimated by 126 (837-711).

A less clear  issue associated with the  design  and
weighting  is  related  to examining  relationships
among  variables.  Unweighted analyses such  as
regressions or correlations should not be used unless
the relationships between the variables are the same
across strata. Unless the relationships are independ-
ent of alkalinity map class (and any factor associated
with the alkalinity  map class strata) unweighted
estimates can be biased, just as unweighted means
or medians and total numbers can be.

In this report, the estimated statistics for regression
and correlation  analyses  and  their  associated
standard errors are presented by stratum; thus, they
are unweighted. Analyses  which combine strata
(e.g., on a subregional level) are  weighted and, as
for strata,  the regression  statistics  are unbiased.
However, the standard errors associated with these
combined estimates are biased and, therefore, are
not presented.

5.1.2.2 Evaluation of Alkalinity Map Classes—The
third  level of  stratification for  the design  was
alkalinity  map class  (Section  2.2.1).  In order to
evaluate the effectiveness of the stratification based
on alkalinity map class, measured ANC values were
compared to  the ranges of alkalinity for each map
class. For the design to be most efficient, the largest
percentage of lakes with ANC < 100 /ueq L"1 should
be observed in Map Class  1, the largest percentage
with ANC from 100-199 fjeq L~1 should be observed
in Map  Class 2 and the largest percentage having
ANC > 200 fjeq L"1 should be observed in Map Class
3. In general, the map classes used in lake selection
were  good estimates of the measured ANC (Table
5-2), but the  intermediate class (100-199 /ueq  L~1)
was less effectively classified by the maps than were
the other two classes.

The maps also can  be evaluated  by examining the
distribution of ANC within  map classes. Table  5-3
gives  the estimated statistics Qi, median (M),  and
Cu for the distribution, F(x). In all  cases, Map Class
1 had lower quintile values than  either Map Class
2 or 3. Map Class 3 had the highest quintile values
of ANC. These results indicate that the map classes
Table 5-2.    Composition of the Alkalinity Map Classes8:
           Numbers and Percentages of Lakes Having
           Measured ANC in Those Same Classes, Western
           Lake Survey - Phase I

                Estimated Percentage:
                Measured ANC (/jeq L~1)
               < 100   100-199  200-400
California (4A)
Map Class 1
Map Class 2
Map Class 3

75.9
66.0
33.3

16.7
17.0
14.3

7.4
17.0
52.4

1727
446
227
Pacific Northwest (4B)
   Map Class 1   62.7     27.1      10.2         586
   Map Class 2   47.2     34.0      18.9         567
   Map Class 3     8.5     31.9      59.6         553

Northern Rockies  (4C)
   Map Class 1   79.2     18.9       1.9         331
   Map Class 2   36.0     34.0      30.0         573
   Map Class 3   22.5     10.0      67.5        1475

Central Rockies (4D)
   Map Class 1   69.8     30.2       0          789
   Map Class 2   40.4     44.7      14.9         928
   Map Class 3   23.1     12.8      64.1         582

Southern Rockies (4E)
   Map Class 1   69.6     23.9       6.5         150
   Map Class 2   26.9     30.8      42.3         235
   Map Class 3     9.8     19.5      70.7        1223
West (4)
Map Class 1
Map Class 2
Map Class 3

72.4
43.9
17.4

21.9
34.6
16.5

5.7
21.6
66.1

3583
2750
4060
aMap Class 1: ANC < 100 ueq L~1 (Omernik and Griffith 1986).
 Map Class 2: ANC 100-199 //eq L"1 (Omernik and Griffith
           1986).
 Map Class 3: ANC 200-400 ^eq L'1 (Omernik and Griffith
           1986).
bN = Estimated number of lakes in the target population in each
     map class.
used in selection led to increased efficiency in the
design.

5.2   Description of Target Population
5.2.7  Number of Lakes Sampled
A total of 973 probability sample lakes was selected
from the  map population. Of those, 95 were classified
as non-target by examination of large-scale maps,
98 were  classified as non-target when visited, and
60 were  not visited (Section 2.2.2). Data from water
samples  collected from 720 lakes were subsequently
considered for use in making population estimates.
One lake which was larger than 2000 ha was
excluded from population estimates (Section 5.2.2);
thus, the number of lakes upon which  population
estimates are based  is 719.

Of the 42 special interest  lakes selected, 32 were
sampled. The data collected from special interest
                                                                           55

-------
Table 5-3.   Population Statistics for ANC by Alkalinity Map
          Class8, Western Lake Survey - Phase I
ANC (neq L~1)

California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
Stratum
4A1
4A2
4A3
4B1
4B2
4B3
4C1
4C2
4C3
4D1
4D2
4D3
4E1
4E2
4E3
O.!b
32.4
44.1
56.9
30.9
33.9
128.6
35.8
69.7
91.7
60.3
62.5
92.3
40.6
63.4
127.4
Mc
54.4
76.7
200.2
82.7
108.8
234.8
50.9
119.8
425.5
87.9
113.9
270.0
74.6
175.7
440.7
Q4d
106.2
165.6
597.0
158.0
195.9
475.3
98.1
406.5
1230.7
108.2
171.3
1239.7
117.6
422.6
1120.5
"Map Class 1: ANC < 100 neq L'1 (Omernik and Griffith 1986).
 Map Class 2: ANC 100-199 A/eq L"1 (Omernik and Griffith
           1986).
 Map Class 3: ANC 200-400 neq L"1 (Omernik and Griffith
           1986).
bQ, = first quintile (20th percentile).
CM  = median (50th percentile).
dO-4 = fourth quintile (80th percentile).

lakes are presented in Volume II and summarized
in Section 5.8.  Because these lakes were  not part
of the random selection process, weighting factors
do not apply in this case, and the representativeness
of these  lakes  with  respect  to the  chemical
characteristics of the lake population as  a whole is
uncertain.

5.2.2  Treatment of Shallow, Thermally
        Stratified, and Large Lakes

In the WLS-I, as in the ELS-I, the intent of the design
was to  sample lakes in a sufficiently deep (> 1.5
m) location when they were isothermal (i.e., mixed).
This sampling objective was not achieved in all cases;
consequently, it  was necessary to  evaluate  the
potential influence of samples collected from shallow
and stratified lakes on the population estimates. The
number of stratified lakes sampled was small (Table
5-4) and thus their impact on the overall  population
estimates is minimal.

Although there  are apparent differences between
shallow lakes  and  deep  lakes  with respect to
chemistry (Table 5-4), the population estimates for
the  shallow  lakes  are  based  on an  insufficient
number of sample lakes (n*** = 7) from which to
draw any inferences. The small number  of shallow
lakes had little influence on  these population
estimates (Table 5-4) so shallow lakes were not
excluded from population estimates.
Concentrations of ANC, calcium, and sulfate (with
the exception  of the median value) are  higher in
stratified lakes than in  mixed lakes. However, most
of the stratified lakes in the West were low elevation
reservoirs. Thus, it is inappropriate to compare high
elevation, mixed lakes with low elevation reservoirs
to assess whether  stratification influences lake
chemistry. For example, a comparison of unstratified
lakes and  reservoirs to  unstratified reservoirs
indicates much larger differences in chemistry than
observed between stratified  and  unstratified lakes.
Therefore,  there appears to be  little basis for
designating stratified  lakes as a separate class in
the data presentations.

Lakes > 2000 ha in surface area were treated as
a separate subpopulation in the ELS-I data analysis
(Linthurst et al. 1986) to minimize the influence of
a very few  large  lakes on the cumulative area!
distributions  [G(x)].  To  maintain parallel data
presentations with  the  ELS-I, the  population
estimates for the WLS-I also exclude lakes > 2000
ha. However, only one lake > 2000 ha was sampled
in the WLS-I (Lake Almanor, California, ID number
= 4A3-017),  so eliminating it had little influence on
the results for F(x) or G(x). Data for Lake Almanor
are retained in the  data  base  and are shown in
Volume II. Lakes  < 4  ha were retained as part of
the target population. The relationship between lake
area and water chemistry is discussed in Section
6.4.1.

5.2.3  Description of Target Population and
Sample

Table 5-5 gives the components of the population
and  sample  by strata, subregions, and region.  All
estimates were made with the equations in Section
2.3. The weights provided in Table 5-5 are approp-
riate for all analyses in which weighting is necessary.

5.3   Physical  Characteristics of Lake
Populations
Marked differences in  lake elevation among  subre-
gions  were  observed (Table  5-6).  The lowest
elevation lakes were in the Pacific Northwest (4B),
where the Q4 value for lake elevation (1710m) was
lower than the Qi values  for all  the other western
subregions except the Northern Rockies  (4C, Qi =
1660 m). The Southern Rockies (4E) had the highest
median and Q4 lake elevations (3264 and 3552 m,
respectively),  although these  values were not
markedly higher than those in the Central Rockies
(4D) or California (4A). The Central Rockies had the
least variation in  lake elevation (as indicated by Qd
= OU - Qi), and California had the most.

Differences  in watershed  slope among subregions
were not as marked as elevational differences and
                       56

-------
Table 5-4.    Population Statistics for ANC, Calcium, and Sulfate in Shallow, Deep, Stratified, and Unstratified Lakes, Western Lake
             Survey - Phase I
ANC (MCI
Subpopulations of Lakes3 n'** N
All lakes 719 10393
Shallow « 1.5m) 7 79
Deep {> 1.5m) 712 10314
Stratified 36 459
Unstratified 683 9934
Unstratified Reservoirs 40 640
QI
54.5
70.3
54.5
87.0
54.1
279.4
M
:jL-1>

119.4
188.4
117.9
235.4
115.7
631.7
Q4
425.6
212.3
427.9
567.5
420.6
1315.5
Ca+2 (^eq L~1)
QI
42.9
28.4
42.9
78.5
42.2
176.6
M
92.4
49.4
92.5
137.4
89.7
428.0
Q4
354.9
139.2
355.9
398.6
351.4
1045.8
S04~
o,
6.2
7.8
6.2
8.0
6.2
20.9
2 (^eq L !)
M
18
11
19
18
18
48
.9
.2
.2
.2
.9
.6
Q4
38.7
13.6
38.7
49.7
38.0
132.4
n"* = number of lakes from which samples were obtained.
N = estimated number of lakes in the subpopulation.
Q! = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
aLakes > 2000 ha in area are not included.
Table 5-5.    Description of the Target Population9
             Survey - Phase I
Sample, and Weighting Factors Excluding Lakes > 2000 ha. Western Lake
STR
                            n*
                                                     W
                                                                                   SE(N)
                                                                  SE(A)
4A1
4A2
4A3a
4B1
4B2
4B3
4C1
4C2
4C3
4D1
4D2
4D3
4E1
4E2
4E3
1885
538
383
695
724
781
343
675
2317
885
1024
1061
150
261
1784
60
65
72
70
70
70
60
60
65
60
60
75
60
63
63
54
53
42
59
53
47
53
50
40
43
47
39
46
52
41
31.978
8.422
5.416
9.929
10.700
11.766
6.246
11.455
36.875
18.356
19.744
14.918
3.261
4.526
29.834
1726.81
446.37
227.47
585.81
567.10
553.00
331.04
572.75
1475.00
789.31
927.97
581.80
150.01
235.35
1223.19
67.58
23.94
20.34
28.87
34.78
42.01
7.85
30.38
139.79
38.89
40.78
60.90
0.00
9.11
106.19
14256
10822
10931
7676
9185
51177
1831
4136
30097
9242
7726
12818
711
1119
10123
2409
3355
3296
1566
3469
23022
196
794
11284
1877
1121
5324
65
136
2979
Subregion
4Aa
4B
4C
4D
4E
Region
4a
STR =
N*
n* =
n*** =
W
N
SE(N) =
A
SE(A) =
2806
2200
3335
2970
2195

13506
stratum.
map population.
number of lakes
number of lakes
weight.
estimated target
standard error of
197
210
185
195
186

973


149
159
143
129
139

719











2400.65
1705.91
2378.79
2299.08
1608.55

10392.98


74.52
61.71
143.27
82.98
106.58

219.39


36009
68038
36065
29786
11953

181851


5284
23335
11314
5756
2983

27248


in the probability sample.
sampled.

population
N.


size.





















estimated area of target population.
Standard error of
A.






aLake Almanor (ID number = 4A3-017) is not included in these estimates (Section 5.2.2). The estimates for the target population
 which included Lake Almanor are:
 STR 4A3: N = 232.89, SE(N) = 20.22, A  =  (65148), SE(A) = 48927.
 Subregion 4A: N = 2406.07, SE(N) =  74.52, A = 90227,  SE(A)  = 49101.
 Region 4: N = 10,398.40, SE(N) = 219.38, A = 236069, SE(A)  = 55906.
                                                                                         57

-------
Table 5-6. Population Statistics for Physical Variables for Lakes in the Target Population.
Lake Elevation (m)

California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West

California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West

4A
4B
4C
4D
4E
4

4A
4B
4C
4D
4E
4
QI
2136
889
1660
2804
2659
1746

QI
2.3
2.3
2.2
3.0
1.5
2.2
M
3008
1437
2064
3040
3264
2613
Lake Area (ha)
M
4.6
4.5
4.6
5.4
3.5
4.6
Q4
3330
1710
2454
3255
3552
3237

Q4
12.3
18.3
8.9
15.7
7.1
11.7
Watershed Area (ha)
QI
47
36
43
67
35
45

QI
4.9
3.3
3.5
4.8
2.8
3.8
M
145
98
95
178
75
118
Site Depth
M
8.8
7.6
8.8
8.9
5.1
7.8
Q4
440
381
214
618
403
436
(m)
Q4
20.0
21.9
17.3
19.0
10.4
17.7
Western Lake Survey -
Phase 1
Watershed Area to
Lake Area Ratio
QI
11.8
9.6
10.8
13.1
11.4
11.0
M
22.9
19.8
23.2
30.2
23.5
23.7
Q4
66.7
42.7
48.3
72.3
73.3
58.5
Watershed Slope (%)
QI
30.1
14.3
23.5
17.7
15.5
19.3
M
51.6
37.1
46.3
30.7
31.1
39.9
Q4
62.7
62.9
61.5
49.9
56.9
61.4
Q, = first quintile (20th percentile).
M  = median (50th percentile).
Q4 = fourth quintile (80th percentile).
did not show the same pattern  (Table 5-6).  The
subregions with the highest median (M) elevations,
the Central and Southern Rockies, had the lowest
median watershed  slopes (30.7% and 31.1%,
respectively). California had  the highest median
watershed slope (51.6%) and the Pacific Northwest
had the greatest variability in watershed slope (Qd
= 48.6%).

The  largest variation in watershed size and the
largest watersheds (at Qi, M, and Q4) were found
in the Central Rockies (M = 178 ha. Table 5-6). The
median watershed area in California (145 ha)  was
also considerably larger than in other subregions.
The  Southern Rockies  had the  smallest median
watershed  area (75 ha) which varied considerably
(Qd = 368  ha). Watershed area  in  the Northern
Rockies was considerably less variable than in other
subregions (Qd = 171  ha).

The smallest watershed area  to lake area ratio, as
well as the  least variability (Qd =  33.1)  in  this
parameter, occurred  in the Pacific Northwest (M =
19.8, Table 5-6). Variability was also  low  in the
Northern Rockies  (Qd = 37.5). The Southern Rockies
had the greatest variability in this  ratio. The Central
Rockies had the highest median watershed area to
lake area ratio (30.2) and a variability similar to that
in California.

The  Southern Rockies had the smallest lakes (M =.
3.5 ha) and the least variation (Qd) in lake area (Table
5-6). Median and Qi values for lake size were similar
among  the other subregions, but Q4 values were
much higher in the Pacific Northwest and the Central
Rockies than in California and the Northern Rockies.
The  shallowest lakes occurred in the  Southern
Rockies (M = 5.1 m, Table 5-6). Site depth was similar
among the other subregions (median values from
7.6 to 8.9 m).

Drainage lakes were predominant in all subregions,
ranging from 58.8 percent in the Southern Rockies
to 83.8 percent  in  California (Table 5-7).  The
percentage  of seepage  lakes  ranged from 24.0
percent in the Central Rockies to  11.8 percent in
California. Seepage lakes were second in importance
in all subregions except in the Southern  Rockies
where  reservoirs (21.0%)  were  slightly more
common.  Reservoirs comprised 3.8 to 4.7 percent
of the lakes  in each of the other subregions. Closed
lakes comprised 0.3 to 3.4 percent of the lakes in
each subregion.

In the analysis of land use types, the  categories of
Anderson et al. (1976) were used (Section 2.4.5).

Table 5-7.    Population Estimates of the Percentage of Lakes
            in Categories by Lake Type. Western  Lake
            Survey - Phase I
Hydrologic
Drainage
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West
4A
4B
4C
4D
4E
4
83.
71.
80,
69.
58,
74.
,8
8
,5
,3
,8
.0
Seepage
11.
22.
12,
.8
2
,3
24.0
20.0
17.6
Lake
Type1
Reservoir
4.
4.
3.
4,
21.
6,
.0
,7
.8
,7
.0
.9
a

Closed
0.
1.
3,
2
0
1
.4
,3
.4
.1
.3
.6
aWetzel (1983).
                       58

-------
Evergreen forest and tundra were the predominant
land cover types  in the watersheds of the WLS-I
(Table 5-8). Evergreen forests were most important
in the Pacific Northwest and in the Northern Rockies,
where they  dominated  77.8 percent  and  52.3
percent, respectively, of the watersheds. Tundra was
the predominant land cover in 57.1  percent of the
watersheds in the Central Rockies and 54.0 percent
of the watersheds  in both  California  and  the
Southern  Rockies. Barren land was  the only other
type of land  cover to dominate over 5 percent of
the watersheds in any subregion. Barren  land was
predominant  in 25.5 percent  of the  watersheds in
the Northern  Rockies, although this  unusually high
percentage may be an artifact of using different land
use information sources for most of this subregion
because USGS maps were not available.

5.4  Population  Estimates
for Primary Variables

Population estimates for  six key  variables are
presented below. These were selected for discussion
because of their  direct relevance to the potential
effects of  acidic deposition.  All values cited are the
estimated population statistics, but are not always
indicated  as  estimates in  the text.  The reference
values are consistent with those in the ELS-I report
(Linthurst  et  al.  1986) and  were  not  based on
inspection of WLS-I  data.  These  reference values
were  selected to  facilitate comparisons between
subregions and are not meant to imply critical or
background  conditions. Comparisons  between
subregions in the Eastern and Western Lake Surveys
can also be made using the estimates for lakes given
the reference values.

5.4.1  Acid Neutralizing Capacity

5.4.1.1   Reference  values—Reference values for
ANC were selected to discuss differences among
subregions. Lakes with ANC  <0/ueq L~1 were defined
as acidic.  A  reference value of  50 /ueq L~1  was
selected  to  represent "low" acid neutralizing
                         capacity. Lakes with ANC < 50 /ueq L 1, while having
                         some buffering capacity, have been demonstrated
                         to have decreases in  ANC during  snowmelt  (Alt-
                         shuller and  Linthurst 1984), although  snowmelt
                         runoff has been reported to have also caused loss
                         of ANC in lakes > 50 /ueq L"1. An ANC of < 200
                         fjeq L~1 has been  used frequently as a criterion to
                         separate  lakes  that may be  "sensitive" to acidic
                         deposition from lakes that may  not  be (Swedish
                         Ministry of Agriculture  1982, Altshuller  and
                         Linthurst 1984).

                         It is important to  note that the estimates for  ANC
                         < 200/ueq L"1 in Table 5-9 are cumulative and include
                         lakes with ANC < 50 /ueq L~1. Only one  lake (Fern
                         Lake, Wyoming, ID number = 4D3-017) in the West
                         had ANC < 0  /ueq L~1; thus, no  summary  table is
                         provided for this reference value, whereas Linthurst
                         et al. (1986) used this value to present results for
                         lakes  in  the  eastern  United States.  Cumulative
                         frequency and  areal  distributions  for ANC are given
                         in Section 5.5 and Volume II.
                         5.4.1.2  Subregional Estimates and Distribution of
                         Lakes Having  Low Acid Neutralizing Capacity—
                         California (4A) contained  the highest  estimated
                         percentage and number of lakes (36.7%, 800) in the
                         West with ANC < 50 /ueq L"1 (Table 5-9). The next
                         highest percentage (19.5%) occurred in the Pacific
                         Northwest (4B), followed by  12.7  percent in the
                         Northern Rockies  (4C). Less than seven  percent of
                         the lakes in the Central Rockies (4D) and Southern
                         Rockies (4E) had ANC < 50 /ueq L"1.
                         California also had the highest percentage  of lakes
                         (86.6%) with ANC < 200 yueq L"1 (Table 5-9). The
                         Central Rockies had the next highest percentage
                         (77.8%), followed by the Pacific Northwest with  71.6
                         percent. Slightly more  than one-half of the  lakes in
                         the Northern Rockies and 39.4 percent of the lakes
                         in the Southern Rockies had ANC  < 200 /ueq L"1.

                        Lakes with ANC < 50 /ueq L"1 were found throughout
                        California, particularly in the southern Sierra  Nevada
                        (Figure 5-2). In the Pacific Northwest, lakes with ANC
                        < 50 //eq  L"1 were  located along the Cascade Crest
Table 5-8.   Population Estimates of the Percentage of Watersheds Dominated8 by Land Use/Land Cover Categories,
          Western Lake Survey - Phase I

                                                Land Use/Land Cover Categories'3

                       Agriculture  Range  Deciduous Evergreen   Mixed    Barren    Tundra    Snow
West
0.1
1.9
0.6
46.9
0.6
                                                                   6.7
38.4
                                                         0.4
                                                                 Total
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
4A
4B
4C
4D
4E
0
0.7
0
0
0
0.6
0
1.6
3.5
4.0
0
0
0
0
3.7
41.1
77.8
52.3
32.5
35.4
0
3.4
0
0
0
2.6
0
25.5
1.4
0
54.0
10.0
14.3
57.1
54.0
0
2.5
0
0
0
98.3
94.4
93.7
94.5
91.7
                                                         95.6
3 More than 50 percent of the watershed in the category.
bAnderson et al.  (1976).
                                                                         59

-------
Table 5-9.   Population Estimates of Lakes with ANC S 50 or
          < 200 f/eq L~1, Western Lake Survey - Phase I

                        ANC <50 peq L'1
                  PC
                    9c
California      4A 0.367
Pacific Northwest 4B 0.195
Northern Rockies 4C 0.127
Central Rockies  4D 0.069
Southern Rockies 4E 0.046
          880 1081 0.234  842312544
          332  415 0.043  2916 4197
          301  413 0.037  1335 1792
          158  240 0.049  1454 2584
           74   96 0.031   374  509
West
4  0.168  1745 20040.0801450118988
                        ANC <200^eq L~
                       fl.
                                  9c
California      4A 0.866 2078 22330.6122203427706
Pacific Northwest4B 0.716 1222 13330.2161472119624
Northern Rockies 4C 0.507 1205 1411 0.185  6669 8351
Central Rockies  4D 0.778 1788 19290.6591962724327
Southern Rockies 4E 0.394  634  7870.213  2540 3232
West
             4  0.666 6926  7276 0.361 65591 74624
pc  = estimated proportion of lakes with ANC S 50 pieq L 1 or
     <200 neq L~1.
Nc  = estimated number of lakes with ANC < 50 ^eq L~1 or
     < 200 f^eq L~1.
Ncu = 95 percent upper confidence limit for Nc.
gc  = estimated proportion of lake area with ANC < 50 f*eq
     L~1  or < 200 ^eq L~1.
Ac  = estimated area of lakes with ANC < 50 /jeq L~1 or
     <200neq L~1.
Acu = 95 percent upper confidence limit for Ac.
refer only to the areas covered by the Western Lake
Survey - Phase I and conditions  in lakes outside the
area sampled cannot be  inferred.  The  state of
California contained the largest number  of lakes
(880) with  ANC  <  50 yueq L"1.  An  estimated 219
lakes in Washington and 189 lakes in Idaho had ANC
< 50 //eq L~1. The estimated numbers of lakes with
ANC < 50 /ueq L~1  were lowest for Colorado (70)
and Utah (20). Of the eight states for which estimates
were calculated, all  had several  hundred lakes with
ANC < 200 fjeq  L~1. These estimates ranged from
461 lakes in Oregon to 2078 lakes in California.


5.4.2  pH

5.4.2.1  Subregional  Estimates—The two refer-
ence values chosen for pH  in the ELS-I were 5.0
and 6.0. Because only one lake (Fern Lake, Wyoming,
ID number = 4D3-017) had pH < 5.0,  population
estimates are presented here only for lakes with pH
< 6.0. Cumulative frequency and areal distributions
for pH are given in Section 5.5 and Volume II.

Only one percent of  the  lakes in the West were
estimated to have  pH <  6.0 (Table 5-11). In the
Northern (4C) and Southern (4E) Rockies, no lakes
with pH < 6.0 were  sampled. The Pacific Northwest
(4B) had the highest percentage  (2.4%) of lakes with
pH < 6.0. In California (4A) and  the Central Rockies
(4D), an  estimated 1.3 percent of the lakes had pH
<6.0.
in Oregon, throughout the Cascades in Washington,
and  on the Kitsap Peninsula (Figure 5-3).  In the
Northern Rockies, lakes with ANC < 50 /ueq L~1 were
located principally in the Bitterroot Range of western
Montana  and northern Idaho and  in the Sawtooth
Mountains in Idaho (Figure 5-4). Some low ANC lakes
were also found in the Salmon River Mountains, the
Clearwater Mountains, and the Selkirk Mountains
in Idaho, and in the Lewis Range in Montana. Lakes
in the Central  Rockies  with ANC < 50 /ueq L"1
occurred in the Wind River  Range and the Big Horn
Mountains in Wyoming, the Beartooth Mountains
and the  Yellowstone Plateau in  Wyoming  and
Montana, and the Uinta  Mountains in Utah (Figure
5-5). In the  Southern  Rockies, lakes with ANC <
50 /ueq L"1 were found primarily in the  San Juan
Mountains and in the Front Range. Some low ANC
lakes were also found in the Park Range in Colorado
and the  Medicine Bow Range in  Wyoming (Figure
5-6).

5.4.1.3   State Estimates—Population estimates for
lakes with ANC < 50 or < 200 A 50  //eq L~1 Table 5-12). The
                                      highest percentage and number (33.7%, 543) of lakes
                                      with SOT2 > 50 jueq L"1 occurred in the Southern
                                      Rockies. The  Central  Rockies was estimated  to
                       60

-------
Figure 5-2.    Locations of lakes sampled in California (4A) with ANC (/jeq i"V in  one of three classes ( • .' ANC < 50.
               -\- : ANC > 50-200. and  O  : ANC > 200). Western Lake Survey  - Phase I. Geographic area covered by
              the target population is bounded by a dashed line.
          Acid Neutralizing
          Capacity (/jeq L'^)

          •   <50

          +   > 50-200

          O   >200
                                                                                       61

-------
Figure  5-3.    Locations of lakes sampled in the Pacific Northwest (4B) with ANC (pet/ L'1) in one of three classes ( • :
              ANC < 50.  -+- : ANC > 50-200. and  O : ANC > 200). Western  Lake  Survey - Phase I. Geographic area
              covered by the target population is bounded by a dashed line.
                          CT--

\ /
«&?!?
W°
vt 1 •> +
^V «NV I ' •
^Kr °:
W^if , o
IJ\I \ + °r>
•//* W f +

fo£?K
^/fr*H '
v4f _y/ /
WV/ /o
1 / •
/ +
•
/' "' °
v 	 \
/
1--
t
/
V t'
\ '' + ••
\ '*
1 »
I t
1 >• __
V 	 "-ix^^
— ^-^r
/*+*
,- ?
t
,< +
/'•f
/
/ •»• +
/ 0 *
1 -t-
< I--
1 •
-t-
' _
' „ +
x- o T^
/ +
|
+*°
. T
»*
/ ^«1
"
I fr /
•> /
"^
/
i' ^J
^ \
1 •*- '
/ + '
\o \
x y
\ /•
' 	 • — T 	 T 	 : 	
o • °? o i
o- 0° -. ',
+ . V > l
+ B. + Ov -
4 + + o o X^
' ^^ +\
+ + 4 )
o. + + «'
• ••• ^
j.T /
_i. f 9 '
•'Wf O '
**• + v
\+ 0 ^
y-«., i
\
X
N
^
0 'l
+ t'
"^ \
o o /
(*/ WX>
0 '-,
** ^ ^
"~ \
I "— — - >. 	
1 _S*~*'~~~^
/ **•
-i-~^ ^^^^^
1 >«y^-.X^
(
i Off
/
/
1
1
1
j
/
;






\
\
1

/

, Acid Neutralizing
Capacity f/ieq L'^)
m <50
+ > 50-200
0 >200



























\
\
1
— \











/
/
/
/
(
V
^




















                           62

-------
Figure 5-4.    Locations of lakes sampled in the Northern Rockies (4C) with ANC (peq O in one of three classes ( • : ANC
              < BO, -\- : ANC > 50-200, and  O -' ANC > 200).  Western Lake Survey - Phase I, Geographic area covered
              by the target population is bounded by a dashed line.
    4C

/ o
I

\ o
v"v
s ^ 'v

^ (

1
/
/
\
V _







^
0




0



o \
o ° \

^.
{ "*». o \
0 V ' + V
<'•*)' x

1 ^**' ^ O X

\ ^ o x
\ / \
r, ' \
\ * " ' '

i \ ^ i '
1 \ > » 0 '
' J . ' ' O ' «*T
                      WA
                                                                             Acid Neutralizin
                                                                             Capacity djeqL
                                                                                <50
                o/?
                                                                              + > 50-200
                                                                              °>200
                                                                                     63

-------
Figure 5-5.    Locations of lakes sampled in the Central Rockies (4D) with ANC f/jeg L V /" one of three classes I  • / ANC
              < 50. -f- -' ANC > 50-200. and  O  : ANC > 200). Western Lake Survey - Phase I. Geographic area covered
              by the target population is bounded by a dashed line.
                                                        4D
                                                                     \++1
                                                                                         MT
                                                                                         WY
                                                                        Acid Neutralizing
                                                                        Capacity (tteq L~*)

                                                                        •   <50

                                                                        -|_   > 50-200

                                                                              >200
                           64

-------
Figure 5-6.   Locations of lakes sampled in the Southern Rockies (4E) with ANC (peg L ^ in one of three classes ( • : ANC
           < 50.  -\- : ANC > 50-200. and O -- ANC > 200). Western Lake Survey - Phase I. Geographic area covered
           by the target population is bounded by a dashed line.
                                               /  /
                                                       WY
                                                                   4E
                           '~
                                                            CO
                                I         O
                                I  O
                   ..J^S     °  f/
                          S;"~"   °    v>/  -
                          O,    *    r*   x-'  >
                             -1   +     .   \        i
                                  o  o
                                            \^
                                       °   C"iN.
"] \ ' „ \ \ ""•«, Acid Neutralizing
\ ^J t~- \ °\ ' \ Capacity (/jeqL~'}
l' ' ^^ ,' . ^Co
V^'' 10 i Vf" " ^°
^0 - \_ + > 50-200
•x Aft O . ^^ ! .. \
f» ++ "> ^o,"» i O >20°
x T + > \
	 ^x 0 ^ / 1 ,
^x + , it
x +l / /
N T \ ^j
"• '^ '.
NM / , p -N \ ,
/ ^ » ; 'V
_ j

_






                                                                      65

-------
Table 5-10.   Population Estimates and Population Statistics by State for ANC and pH, Western Lake Survey - Phase I. These
            Estimates and Statistics Refer Only to Those Lakes Within the Study Area8
State"
                                         Estimated Number of Lakes (UCL)C
                                                                                    Population Statistics
                                                                                      ANC (/jeq
                              (N)
                                       ANC < 50
                                                          ANC< 200
                                                                         -1
M
California 147 2390 880(1081)
Colorado 132 1476 70 (90)
Idaho 72 972 189 (284)
Montana 80 1597 160 (240)
Oregon 55 551 113 (165)
Utah 30 548 20 (51)
Washington 117 1338 219 (290)
Wyoming 83 1480 94 (158)

Stateb n*** (N) pH < 6.0
California 147 2390 32(84)
Colorado 132 1476 O(-)
Idaho 72 972 O(-)
Montana 80 1597 O(-)
Oregon 55 551 10(25)
Utah 30 548 O(-)
Washington 117 1338 31(59)
Wyoming 83 1480 30(63)
n*" = number of lakes from which samples were obtained.
N = estimated number of lakes in the area of the state covered by
Q, = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
(-) = undefined.
a Refer to Figure 2-1 (Section 2.1.3) for study area.
bOnly one lake was sampled in New Mexico and two in Nevada; thus.
2078 (2233)
591 (739)
599 (727)
824(1035)
461 (558)
484 (620)
822 (935)
1068(1234)











WLS-I.





estimates for these
34.4
114.8
49.8
79.0
44.7
59.9
55.3
75.0

QI
6.6
7.1
6.8
7.0
6.6
6.8
6.7
6.9







states were not
62.5
305.1
108.3
191.9
104.0
85.8
156.9
128.4
pH
M
6.9
7.6
7.0
7.4
7.1
7.0
7.0
7.2







computed.
136.6
981.8
614.4
977.6
198.3
114.0
461.8
241.6

0-4
7.4
8.4
8.0
8.0
7.3
7.3
7.5
7.5








c The 95 percent upper confidence limit (UCL = Nco) is shown in parentheses
Table 5-11.   Population Estimates of Lakes with pH < 6.0,
            Western Lake Survey - Phase I
                                                        Table 5-12.     Population  Estimates of Lakes with Sulfate
                                                                      >50 fjeq \-~\ Western Lake Survey - Phase I
                    PC
                                                "cu
                                                                            PC
                                                                                             9c
California 4A
Pacific Northwest 4B
Northern Rockies 4C
Central Rockies 4D
Southern Rockies 4E
0.013
0.024
0.000
0.013
0.000
32
41
0
30
0
84
73
—
63
-
0.002
0.002
0.000
0.022
0.000
64
106
0
647
0
168
193
—
1572
-
California 4A
Pacific Northwest 4B
Northern Rockies 4C
Central Rockies 4D
Southern Rockies 4E
0.078
0.171
0.107
0.056
0.337
187
292
256
128
543
300 0.093
374 0.138
399 0.414
198 0.056
711 0.569
3363 6793
9389 17408
14914 30241
1667 3123
6805 11737
West
               4  0.010  103   172   0.004   817  1752
pc  = estimated proportion of lakes with pH 5= 6.0.
Nc  = estimated number of lakes with pH 2 6.0.
Ncu = 95 percent upper confidence limit for Nc.
gc  = estimated proportion of lake area with pH < 6.0.
Ac  = estimated area of lakes with pH ^ 6.0.
Acu = 95 percent upper confidence limit for Ac.
(-)  = undefined.

contain the fewest lakes (128) and lowest percentage
(5.6%) of lakes with S04"2 > 50 //eq  L"1.

5.4.4  Calcium

A reference value for calcium of 50  //eq L"1 selected
in the ELS-I  was also   used  to  compare the  five
western subregions. Cumulative frequency and areal
distributions for Ca+2 are given in Section 5.5  and
Volume II.
                                                        West
                                                                       4  0.135  14051676 0.1993613854507
                                                        pc  = estimated proportion of lakes with sulfate ^ 50 ^eq L 1.
                                                        Nc  = estimated number of lakes with sulfate ^ 50 ^eq L~1.
                                                        Ncu = 95 percent upper confidence limit for Nc.
                                                        gc  = estimated proportion of lake area with sulfate > 50 ^eq
                                                        In the West, an estimated 24.6 percent of the lakes
                                                        had Ca+z < 50 peq L"1 (Table 5-13). The percentages
                                                        of  lakes with  low levels of Ca+2 varied markedly
                                                        among subregions.  More than one-half (56.5%) of
                                                        the lakes in California were estimated to have Ca+2
                                                        < 50 /jeq L"1 (Table 5-13). Of the target populations
                                                        of lakes in the  Central Rockies and in the Southern
                                                        Rockies, 6.9 and 3.4 percent, respectively, had Ca+2
                         66

-------
5.4.5  Extractable Aluminum, Clearwater Lakes

Population estimates for extractable aluminum are
given for  clearwater lakes only,  which comprised
96.1 percent of the western target population. The
reference value used was 50 /JQ L~1 based on ELS-I
results. Clearwater lakes were defined as having true
color < 30 PCU. Darkwater lakes (> 30 PCD) were
assumed to  have adequate  concentrations  of
dissolved  organic  material to  produce  organic
complexes of aluminum, which can ameliorate the
toxic effects of aluminum (Driscoll et al. 1980; Baker
and Schofield 1 982). Population estimates are given
for extractable Al > 50 //g L~1. Inverse cumulative
frequency and  areal  distributions are given  in
Volume II.

Less than one percent of the  clearwater  lakes in
the West had extractable Al > 50 jjg L~1 (Table 5-14).
Eleven clearwater lakes in the Pacific Northwest and
five  clearwater lakes  in  California had extractable
Al > 50 fjg L~1. No clearwater lakes with extractable
Al > 50 //g L~1 were sampled in the Rockies. These
results sharply contrast those for the northeastern
United States (Section  8).

5.4.6  Dissolved Organic Carbon

Two values were  used to compare  population
estimates for dissolved organic carbon (DOC) in the
ELS-I,  DOC < 2 mg L~1 (low) and DOC > 6 mg L~1
(high). The latter value was selected because, in lakes
with high DOC >  6   mg L~1, organic  anions  can
contribute appreciably  to  the ion  balance (Oliver et
al. 1983). Cumulative frequency and areal  distribu-
tions for DOC are given in Volume II.

Most of the lakes in the West had low DOC (69.7%,
Table 5-15). The percentage of lakes with low DOC
ranged from 56.5 percent in the  Southern Rockies
to 80.7 percent in California. Very few  lakes in the
West had high DOC. The highest  estimated number
of lakes with DOC > 6 mg L~1 was observed in the
Northern  Rockies (196)  followed by the Southern
Rockies (158).
5.5  Statistics for Population
Descriptions, Primary Variables
Direct  qualitative  comparisons  of  lake  chemistry
among subregions can be  made by superimposing
cumulative frequency [F(x>] curves. This approach
can  serve  to highlight  major differences  in  the
distributions of physical or  chemical characteristics
of lakes  among subregions. Superimposed curves
can also show subtle differences among subregional
distributions which are not always evident from an
analysis  of population statistics,  such as quintiles
or medians.
Table 5-13.    Population Estimates of Lakes with Calcium
            <50 neq L~1, Western Lake Survey - Phase I
                  PC
                                  9c
California       4A 0.565  1357 15660.3971428719191
Pacific Northwest 4B 0.284   485  578 0.058 3954  5380
Northern Rockies 4C 0.213   506  6490.073 2617  3615
Central Rockies  4D 0.069   158  2400.034 1000  1696
Southern Rockies 4E 0.034    55   740.023  274  398
West
4  0.246 2562  28440.1222213227383
pc = estimated proportion of lakes with calcium < 50 ^eq L 1.
Nc = estimated number of lakes with calcium < 50 ^eq L"1.
Ncu= 95 percent upper confidence limit for Nc.
gc = estimated proportion of lake area with calcium < 50^ieqL~1.
Ac = estimated area of lakes with calcium < 50 neq L"1.
Acu= 95 percent upper confidence limit for Ac.
Table 5-14.    Population Estimates of Clearwater Lakes with
             Extractable Aluminum >50 yueq L 1, Western
             Lake Survey - Phase I
PC
California 4A 0.002
Pacific Northwest 4B 0.006
Northern Rockies 4C 0.000
Central Rockies 4D 0.000
Southern Rockies 4E 0.000
West 4 0.002
*c
5
11
0
0
0
16
NCU
14
28
—
—
—
35
9c
0.000
0.000
0.000
0.000
0.000
0.000
*c
10
33
0
0
0
43
Acu
24
85
_
—
—
97
pc  = estimated proportion of lakes with extractable aluminum
      >50yugL"1.
Nc  = estimated number of lakes with extractable aluminum
      >50yugL"1
NCU = 95 percent upper confidence limit for Nc.
gc  = estimated proportion of lake area with extractable
      aluminum >50/ug L'1.
Ac  = estimated area of lakes with extractable aluminum
      >50yug L"1
Acu = 95 percent upper confidence limit for Ac.
(-)  = undefined.
The median, Qi and Q4 values also can be used to
compare the distributions among subregions. The
Qi value is useful in comparing variables for which
low concentrations are of interest, such as ANC or
pH.  Similarly, Q4 values are useful  in comparing
variables  for which  high  concentrations  are of
primary interest, such as S04 2 or extractable Al.

The superimposed  cumulative frequency distribu-
tions for  ANC  for the  five western subregions
indicates  that ANC  concentrations  in  lakes in
California  were generally  lower  than  in  other
subregions (Figure 5-7). The distribution for ANC in
the Pacific  Northwest was similar to that in the
Central Rockies. The distributions  indicate  that
generally, the ANC values for lakes in the Northern
Rockies and in the Southern Rockies were substan-
tially higher than in the other subregions.

Comparisons of the distribution curves for pH (Figure
5-7) yield results similar  to those observed for ANC
                                                                           67

-------
Table 5-15.    Population Estimates of Lakes with Dissolved
             Organic Carbon < 2 mg L~1 or >6 /veq L~1,
             Western Lake Survey - Phase I

                  Dissolved Organic Carbon <2 mg L~1
 California       4A 0.807 1937 2105 0.937  33728  42462
 Pacific Northwest 4B 0.681 1161 12770.880  59873  98049
 Northern Rockies 4C 0.708 1685 1929 0.511  18427  29879
 Central Rockies  4D 0.673 1 548 1717 0.772  23005  32426
 Southern Rockies 4E 0.565  90810880.525   6276  10417
 West
              4  0.697 7239 7642 0.777 141309 183389
                  Dissolved Organic Carbon >6 mg L
                  PC
9c
A    A
"c   "cu
California 4A
Pacific Northwest 4B
Northern Rockies 4C
Central Rockies 4D
Southern Rockies 4E
0,
0
0,
0
0
,030
.027
.082
.040
.098
71
45
196
93
158
129
80
326
152
263
0
0
0
0.
0,
.004
.003
.039
.034
,131
156
214
1423
1025
1566
276
400
2481
2065
3660
West
              4  0.054   563  753 0.024 4385 6961
 pc  = estimated proportion of lakes with dissolved organic
      carbon <2 mg L~1 or >6 mg L~1.
 Nc  = estimated number of lakes with dissolved organic
      carbon <2 mg L~' or >6 mg L~1.
 Ncu = 95 percent upper confidence limit for Nc.
 gc  = estimated proportion of lake area with dissolved organic
      carbon < 2 mg L~1 or 2- 6 mg L'1.
 Ac  = estimated area of lakes with dissolved organic carbon
      <2 mg L"1 or >6 mg L"1.
 Acu = 95 percent upper confidence limit for Ac.


 for the three subregions in the Rocky Mountains.
 In contrast to ANC, however, the distributions for
 pH in California and the Pacific  Northwest were
 virtually indistinguishable.

 Sulfate data  are presented as the  inverse distribu-
 tion, 1 -F(x) (Figure 5-8). Higher values of SO4~2 were
 observed in lakes in the Southern  Rockies than in
 other subregions. Lakes in California were generally
 characterized by very low SO*~2 concentrations. The
 distributions were very similar  for  the Pacific
 Northwest and the Northern Rockies, particularly at
 low concentrations. The distribution in the Central
 Rockies was shifted toward higher concentrations,
 which approached those observed in the Southern
 Rockies.

 In most lakes Ca+2 is the dominant base cation; thus
 the distributions for Ca+2 would be expected to be
 similar to those for ANC. For lakes sampled  in the
 West, the  patterns for  Ca+2  distributions closely
 follow those  for ANC distributions (Figure 5-8). The
 relationship between ANC and Ca+2 is  discussed in
 more detail in Section 6.2.

 The distribution curves for  the other  two  primary
 variables, DOC and extractable Al, are not presented
 as superimposed F(x) curves because concentrations
were very low and the differences among subre-
gional curves are indistinguishable. Distributions for
these variables are given in Volume II.

California had the  lowest values  of  ANC at Qi,
median and  Q4 levels (Table 5-16).  The  Pacific
Northwest had the second lowest Qi value, whereas
the Central Rockies had the second lowest median
value of ANC. The Southern Rockies had the highest
estimated quintile values.

The  lowest Qi value  for  S04~2 was estimated for
the Pacific Northwest, whereas  California  had the
lowest estimated median value. Sulfate was highest
at the Qi level  in the Central Rockies and the highest
median value was  estimated for  the Southern
Rockies.

The  distribution for  Ca+2 paralleled that for ANC,
because calcium is the primary base cation (Section
6.2). Calcium  was lowest in California and highest
in the Southern Rockies. The Pacific Northwest had
the second lowest Qi value for Ca*2, whereas the
Central Rockies had a slightly lower median than
did the Pacific Northwest.

Little interquintile difference was observed in DOC
values. Dissolved organic  carbon was generally
lowest in California and  highest  in the Southern
Rockies. Extractable Al  values  were  all low  and
median values  were  near or below the system
decision limit (4 //g L"\ Section 4.2).


5.6   Statistics for Population
Descriptions, Secondary Variables

Many variables measured in the WLS-I, in addition
to the primary variables discussed above, are useful
in characterizing lakes. In this section, related sets
of variables are grouped:  nutrients (nitrate, ammo-
nium and total  phosphorus),  those  related to
transparency (true color,  turbidity and Secchi  disk
transparency), major cations (potassium, sodium and
magnesium),  metals  (iron,  manganese and  total
aluminum), and others (silica, dissolved inorganic
carbon, chloride, conductance and bicarbonate).
                  5.5.7   Nutrients

                  Nutrient  concentrations in western lakes  were
                  generally low.  Estimated Qi  values  for  total
                  phosphorus were less than 2 fig L"1 for California,
                  the Pacific Northwest, and the Northern Rockies, and
                  were less than 4 /ug L~1 for all  subregions (Table
                  5-17).  Median values of total phosphorus were less
                  than 4 /ug L~1 and Q4 values were less than 10 //g L'1
                  for lakes in California, the Pacific Northwest, and
                  the Northern Rockies. The Southern Rockies had the
                       68

-------
Figure 5-7.     Cumulative frequency distributions [F(x)J for ANC (peg L V and pH in California (	1, the Pacific Northwest
               (	), the Northern Rockies (	;, the Central Rockies (	). and the Southern Rockies (	). Western
               Lake Survey - Phase I.
           FM
                 1.0
                 0.8
                 0.6
0.4
0.2
                 0.0
                                                                       	California (4A)
                                                                       	Pacific Northwest (4B)
                                                                       	Northern Rockies (4CJ
                                                                       	Central Rockies (4D)
                                                                       	Southern Rockies I4E)
                 1.0



                 0.8



                 0.6



                 0.4



                 0.2


                 0.0
                      4.0
                                                                         600
                                                                        1	
                                                                         soo
                                                                                                            ;ooo
                                                           ANC (fjeq L"1)
                   California (4A)
                   Pacific Northwest (4B)
                   Northern Rockies (4C)
                   Central Rockies (4D)
                   Southern Rockies (4E!
                                                                PH
                                                                                          69

-------
Figure 5-8.
Cumulative frequency distributions [F(x)J for calcium lueq L'^) and inverse cumulative frequency distributions
[1-F(x)J for sulfate (fteq £"V '" California (	;, the Pacific Northwest (	). the Northern Rockies f	).
the Central Rockies (	), and the Southern Rockies (	). Western Lake Survey - Phase I.
                   1.0
                   o.s -
                   0.6
                   0.4  -
                   0.2  ~
                   0.0
                                                                California I4A)
                                                                Pacific Northwest /4B>
                                                                Northern Rockies I4C)
                                                                Central Rockies (4D)
                                                                Southern Rockies f4E)
                                         200
                                                                          600
                                                                            800
                                                                                                          1000
                                                           Calcium (ueq L'
                                                                          	  California (4A)
                                                                          —  Pacific Northwest (4BJ
                                                                          	Northern Rockies (4C)
                                                                          	Central Rockies (4D)
                                                                          	Southern Rockies (4EI
                    0.0
                                                         40              60

                                                            Sulfate (ueq L~')
                                                                                           80
                                                                                             100
                             70

-------
Table 5-16.    Population Statistics for the Primary Variables, Western Lake Survey - Phase I


California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West


California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies


4A
4B
4C
4D
4E
4


4A
4B
4C
4D
4E

0-1
34.5
51.7
69.1
64.7
113.9
54.5

QI
22.8
37.3
48.8
65.4
79.2
ANC (fiec
M
62.6
139.8
195.2
104.9
317.0
119.4
Ca+2 (jieq I
M
43.2
95.8
154.2
92.7
233.1
|L-1)
a
136
293
1011,
207
1039
425
--')
Q<
106
206
730.
169,
724,

i
.9
.1
.4
.1
.4
.6

t
.1
.0
,0
,3
.7

QI
6.59
6.63
6.90
6.91
7.10
6.81

QI
1.2
2.1
1.1
0.0
0.1
pH
M
6.94
7.00
7.35
7.18
7.60
7.16
Ext Al (Mg L'1)a
M
2.7
5.6
3.3
1.1
2.5

Q4
7.38
7.34
7.99
7.47
8.27
7.69

Q4
4.6
11.1
7.2
3.8
6.7

QI
3.6
2.3
6.3
19.1
16.6
6.2

QI
0.46
0.52
0.75
0.68
0.82
S04 2 (jieq L
M
6.6
14.7
15.6
24.4
34.6
18.9
DOC (mg L"1)
M
0.97
1.30
1.20
1.25
1.48
')
Q4
13.5
43.3
32.3
35.0
115.5
38.7

Q4
1.93
2.57
2.63
3.23
3.72
West
42.9
                               92.4
354.9
                                                   0.6
                                      2.8
6.6
0.64
                                                                                       1.21
                                                                                                2.76
Q, = first quintile (20th percentile).
M  = median (50th percentile).
Q4 = fourth quintile (80th percentile).
aClearwater (true color < 30 PCU) lakes only.


Table 5-17.    Population Statistics for Secondary Variables (Nitrate, Ammonium and Total Phosphorus), Western Lake Survey -
             Phase I



California
Pacific Northwest
Northern Rockes
Central Rockies
Southern Rockies
West



4A
4B
4C
4D
4E
4


Q,
0.1
0.0
0.1
0.2
0.1
0.1
N03~
(fjeq L'1)
M
0.4
0.1
0.4
0.4
0.5
0.4


CU
2.5
1.4
2.1
3.1
2.5
2.2


0.1
0.0
0.0
0.0
0.0
0.0
0.0
NlV
(fjeq L'1)
M
0.0
0.0
0.0
0.0
0.0
0.0


0.4
0.6
0.6
0.4
0.4
0.6
0.5


Qi
0.7
1.6
1.9
3.0
3.6
1.7
Total P
G«g L"1)
M
2.9
3.5
3.7
6.9
8.1
4.7


O.4
7.3
7.3
9.7
12.6
26.6
10.9
Qi = first quintile (20th percentile).
M = median (50th percentile).
Qt = fourth quintile (80th percentile).
highest concentrations of total phosphorus (Q4 = 26. 6
fjQ  L"1), although the median value was only 8.1
Nitrate and ammonium concentrations were also low
for lakes in all subregions. The Qi values for nitrate
were less than or equal to 0.2 /ueq L"1 and median
values were less  than or equal to  0.5 /ueq  L"1.
Ammonium concentrations at the Q4 value did not
exceed 0.6 /ueq L"1 for lakes in any subregion.

Nitrate may be an important anion during periods
of snowmelt runoff. For  example,  it  has been
regarded as a major contributor to reductions in acid
neutralizing capacity during snowmelt episodes in
northeastern North America (Galloway and  Dillon
1 983). Nitrate levels in western lakes were very low.
The  lakes were sampled during  the fall, well after
the snowmelt period. Biological uptake of nitrate has
                               been documented for  lakes in the North Cascades
                               during open water periods, which could contribute
                               to the  low levels of nitrate found in western lakes
                               (Loranger and  Brakke 1986). The potential  contri-
                               bution of nitrate  to  lakewater acidity during  the
                               spring  or other periods cannot be inferred from the
                               fall sampling.


                               5.6.2  True Color, Turbidity, and Secchi Disk
                               Transparency

                               Lakes  sampled  during the WLS-I were generally
                               clearwater with low DOC (Section 5.4.6) and turbidity
                               (Table 5-18). Consequently, Secchi disk transparency
                               was very high. Lakes in California had very low true
                               color (Q4 = 3.5  PCU) and  very low turbidity (Q4 =
                               0.4  NTU).  Median Secchi disk transparency was
                               7.4 m  and Qi (which corresponds to the  Cu value
                                                                             77

-------
for color and turbidity) was 4.2 m. True color was
similar among the three subregions in the Rockies,
with a somewhat  higher Q4 value in the Southern
Rockies (21.6 PCU). Lakes in the Pacific Northwest
had lower true  color values and, at  the  Q4 level,
higher Secchi disk transparency than the Rockies.
Southern Rockies (Qd = 234.5 /ueq L~1). Magnesium
concentrations in California were low (Q4 = 17.7
/ueq L"1). Lakes in the Pacific Northwest and Central
Rockies had  intermediate levels of magnesium
relative to other western subregions (M = 25.8 and
34.2  /ueq L"1 and  Q4  =  73.4 and 55.2 /ueq L"1
respectively).
5.6.3  Sodium, Potassium, and Magnesium

Values of sodium, potassium and magnesium were
generally quite low (Table 5-19). Although Qi values
for  sodium were similar among all subregions, Q4
values ranged widely (33.6 to 91.7 /ueq L~1). The
lowest sodium concentrations were  observed for
lakes in California. The highest median (33.8 /ueq L"1)
and Q4 (91.7 /ueq L"1) values were estimated for the
Southern Rockies.

Potassium  concentrations  varied  little and were
lowest  in  lakes in California and  the  Pacific
Northwest. Slightly higher concentrations of potas-
sium were observed in the Southern Rockies where
the Q4 value was 21.8 /ueq L~1.

Magnesium concentrations were markedly variable
in the Northern Rockies (Qd  = 234.1  /ueq L~1) and
5.6.4  Iron. Manganese, and Total Aluminum

Concentrations  of  iron,  manganese and total
aluminum were generally low for most lakes and
did not vary  markedly among  subregions (Table
5-20).  Estimated  median values for iron were less
than 18 /ug L"1 for all subregions. The lowest median
value occurred in the Northern Rockies (9.7 /ug L"1).
The maximum estimated Q4 value for iron (66.9
/ug  L"1) was observed  in  the  Southern  Rockies.
Manganese  concentrations were also low;  the Q4
value ranged from 5.3 /ug L"1 in the Central Rockies
to 17.4/ugL~1 in the Pacific Northwest. Median values
for  total aluminum were estimated to be less than
or equal to 30 /ug L"1 for all subregions. The lowest
Qi  and median values for total aluminum were
observed in the Central and Southern Rockies.
Table 5-18.    Population Statistics for Secondary Variables (True Color, Turbidity and Secchi Disk Transparency), Western
             Lake Survey - Phase I


California
Pacific Northwest
Northern Rockes
Central Rockies
Southern Rockies
West


4A
4B
4C
4D
4E
4

Qi
0.0
1.1
40
3.6
3.7
0.0
True Color
(PCU)
M
0.0 3.5
4.2
8.0
8.4
11.3
5.0

CU
0.1
10.1
12.9
17.9
21.6
13.4

Qi
0.2
0.2
0.2
0.3
0.4
0.2
Turbidity
(NTU)
M
0.4
0.3
0.3
0.6
0.7
0.4
Secchi Disk
Transparency
(m)
Q4
4.2
0.7
0.5
1.4
2.0
0.9
Qi
7.4
26
2.4
2.7
1.7
2.6
M
11.1
4.9
4.9
4.2
3.4
4.8
Cu

10.0
7.9
6.9
5.4
8.2
Qi = first quintile (20th percentile).
M = median (50th percentile).
Cu = fourth quintile (80th percentile).


Table 5-19.    Population Statistics for Secondary Variables (Sodium, Potassium, and Magnesium), Western Lake Survey -
             Phase I

                                                                 1)                  Mg*2(Aieq L'1)
                                    M
                                             Cu
  Qi
                                                              M
Q,
                                                                                       M
California
Pacific Northwest
Northern Rockes
Central Rockies
Southern Rockies
West
4A
4B
4C
4D
4E
4
10.1
17.2
11.7
12.4
15.9
12.4
18.7
31.0
24.8
20.8
33.8
23.9
33.6
57.5
50.2
38.6
91 7
49.2
2.7
2.6
2.7
5.2
3.5
29
3.7
4.8
5.2
7.5
9.0
5.6
7.6
10.6
14.8
12.4
21.8
12.8
3.5
11.1
9.3
18.5
21.4
8.9
7.2
25.8
36.8
34.2
91.5
26.4
17.7
73.4
243.4
55.2
255.9
96.7
Qi = first quintile (20th percentile).
M = median (50th percentile).
Cu = fourth quintile (80th percentile).
                       72

-------
5.5.5  Other Secondary Variables

Estimated concentrations of silica were lowest for
the Central Rockies, where the median concentration
was 1.37 mg  L"1  (Table 5-21).  Median Si02
concentrations  in all subregions were less than or
equal to 3.14 mg L~1.

Very low concentrations  of  dissolved  inorganic
carbon (DIG)  were observed for lakes  in California
and in the Central Rockies (Q4= 1.88 and 2.80 mg L"1,
respectively). Median values for DIG in the Northern
and Southern Rockies were only slightly higher than
those in  the  other subregions, but Q4 values were
much higher  (1 2.73 and 1 2.04 mg L"1, respectively).
Chloride concentrations  were lowest in California
and in the Northern Rockies. The highest Q4 values
were observed in the Pacific Northwest, which may
reflect the  proximity of some lakes to the coast
(Sections 6.2.5 and 7.2). The lowest conductance
values  were  estimated  for California (M  = 8.8
/jS cm"1) and the highest  for  the Northern  and
Southern Rockies (M =  25.1 and  37.1  fjS cm"1,
respectively).
                      Calculated bicarbonate concentrations were lowest
                      in California and highest  in the Southern Rockies
                      (Table 5-21). The Pacific Northwest had the second
                      lowest estimated Qi  value,  whereas  the Central
                      Rockies had the  second  lowest median (96.8 /ueq
                      5.7  Statistics for Lakes in Wilderness
                      Areas and National Parks
                      5. 7. 1  Wilderness Area and National Park Lakes
                      The estimated population statistics for ANC, Ca+2 and
                      S04~2 were examined for  wilderness area lakes by
                      subregion and region (Table 5-22). Wilderness area
                      lakes were defined  as  lakes  in  Forest Service
                      Wilderness Areas,  National  Parks,  and the Wind
                      River Roadless Area (Table 2-4). The same relative
                      differences between subregions described in Section
                      5.5 for  the  total target population  of  lakes were
                      observed in the data from wilderness  area  lakes.
                      Wilderness area lakes in  all  subregions had lower
                      ANC values at the median and the first and fourth
                      quintiles than  did non-wilderness lakes. The
                      medians and quintiles for Ca+2 were also generally
                      lower in  wilderness area  lakes but the  differences
Table 5-20.    Population Statistics for Secondary Variables (Iron, Manganese, and Total Aluminum), Western Lake Survey
             Phase I


California
Pacific Northwest
Northern Rockes
Central Rockies
Southern Rockies
West


4A
4B
4C
4D
4E
4

Qi
5.4
5.0
4.6
1.1
5.5
42
Fe
M
14.5
15.7
97
10.8
17.8
137

Cu
37.1
36.1
283
54.9
66.9
431

Qi
0.0
0.0
0.0
0.0
00
0.0
Mn
M
0.0
25
0.0
0.9
1 7
0.8

CU
11.0
17.4
6.6
53
11.4
9.9

Qi
14.4
18.4
16.5
7.5
8.1
13.1
Total Al
(A
-------
Table 5-22.   Population Statistics for ANC, Calcium, and Sulfate in Wilderness Area" Lakes and Non-Wilderness Area Lakes,
            Western Lake Survey - Phase I
                                                           Wilderness Area Lakes


California 4A
Pacific Northwest 4B
Northern Rockies 4C
Central Rockies 4D
Southern Rockies 4E

n
97
90
82
96
90

XV
N
2017
939
1061
1734
807
/
QI
30.6
34.8
54.8
59.1
74.4
<\NC (/jeq L
M
54.5
104.2
94.6
97.5
157.5

QA
106.3
207.8
418.4
173.3
445.3
(
QI
21.1
29.0
38.3
63.8
69.0
Za+z (peq I
M
40.6
74.3
65.4
90.1
121.0
--1)
Q4
82.2
206.9
351.8
141.0
330.4

QI
4.1
3.5
5.8
19.1
14.0
so4-2 (
M
6.9
12.8
10.0
24.2
24.1
:^eq L-1)
Q4
13.7
41.6
21.5
33.2
42.1
West
                    455
6559
46.1
               91.4
               198.8
35.3    71.7
169.7
                                                       5.8  15.4   32.2
                                                         Non-Wilderness Area Lakes
ANC (fieq L~1)
n"" N
California 4A 52 383
Pacific Northwest 4B 69 766
Northern Rockies 4C 61 1318
Central Rockies 4D 33 565
Southern Rockies 4E 49 801
West 4 264 3834
QI
71.8
68.9
106.3
82.4
303.6
94.4
M
163.0
171.5
387.3
140.6
668.2
282.7
n*** = number of wilderness lakes from which samples were
Q4
365.
351.
1227.
552.
1403.
1108.
obtained

9
5
4
5
5
2

Ca + 2 (fjeq L 1)
QI
36.6
50.0
63.4
72.6
220.1
66.6

M
91.4
109.2
336.5
112.3
590.9
201.4

Q4
200.1
205.1
807.5
387.5
1156.3
740.4

Q
1.
2
7.
16
28,
7

,
.4
.0
.9
.5
,0
.7

so4-2 1,
M
3.7
16.2
21.7
27.9
56.2
25.5

.eqL-1)
°4
11.1
47.7
39.1
38.6
198.6
68.8

ft = estimated number of wilderness lakes.
Q, = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).

































'Includes Forest Service wilderness areas, national parks, and the Wind River Roadless Area (Table 2-4)
were  less than those for ANC.  Sulfate values in
wilderness area lakes were lower than  in  non-
wilderness area lakes at the median  and fourth
quintile in all subregions except  California. Sulfate
was lower in non-wilderness area lakes at Qi values
in California, the Pacific Northwest, and the Central
Rockies. These observations indicate that wilderness
area lakes generally contain solute concentrations,
particularly ANC, in the lower  end  of the  range
observed for all lakes in the West.
Another way to examine  differences  between
wilderness area  and non-wilderness area lakes  is
to superimpose cumulative frequency distributions
(Figures 5-9  through  5-11).  Acid  neutralizing
capacity and Ca+2 are consistently lower in wilder-
ness area than in non-wilderness area lakes except
in the  Pacific  Northwest. The differences are
particularly striking in the  Rocky Mountain subre-
gions.  Sulfate is consistently lower  in wilderness
area than in non-wilderness area lakes only in the
Southern Rockies. In the Central Rocky Mountains
and the Pacific Northwest, little difference in S04~2
exists between wilderness area and non-wilderness
area lakes with low concentrations of S04 2- Sulfate
is consistently higher in wilderness area than in non-
wilderness area  lakes in California, although the
differences are not great.
                         5.7.2  Subpopulations in Wilderness Areas and
                         National Parks

                         Acid neutralizing capacity, Ca+2and SO4~2data were
                         examined for  21 Subpopulations  of lakes  in
                         Wilderness Areas  and National Parks (Figure 5-12).
                         The lowest Qi values for ANC were estimated for
                         the Tahoe area in California and the Washington
                         Middle Cascades in the  Pacific Northwest (17.6 and
                         18.0  /ueq L"1, respectively). Other areas having QT
                         values for ANC < 30 /ueq L"1 were the Yosemite area
                         and the Southern Cascades in California, and the
                         Oregon  Middle  Cascades  and  Alpine  Lakes
                         Wilderness in Washington.  As  indicated by the
                         interquintile  difference,  some areas  were very
                         homogeneous, particularly the Tahoe and Yosemite
                         areas  and the Sawtooth and Wasatch/Uinta
                         Mountains. Other areas, such as northwest Montana
                         and northwest Wyoming had large Qd values for
                         ANC. Similar results were observed for Ca+2. Sulfate
                         concentrations were highest and, in general, most
                         variable in areas  of the Central  and  Southern
                         Rockies. In the Southern Cascades in California and
                         the Oregon  Middle Cascades in  the  Pacific
                         Northwest, S04 2 concentrations were very low and
                         uniform. All areas of the  Cascades in Washington
                         had  higher  and  much  more variable  SO4~2
                         concentrations than did areas of the Cascades  in
                         other states.
                       74

-------
Figure 5-9.     Cumulative frequency distributions [F(x)J for
               ANC (fjeq i"V in wilderness area (	) and
               non-wilderness area (•	) lakes in California
               (4A), the Pacific Northwest (4B). the Northern
               Rockies (4C). the Central Rockies (4D). and
               the Southern  Rockies  (4E).  Western Lake
               Survey - Phase I.
          Figure  5-10.     Cumulative frequency distributions [F(x)J for
                          calcium (fjeq i~V in wilderness area (	)
                          and non-wilderness area (•	) lakes in
                          California (4A). the Pacific Northwest (4B),
                          the Northern Rockies (4C), the  Central
                          Rockies (4D), and the Southern Rockies (4E),
                          Western Lake Survey - Phase I.
                                    California (4Aj
                                     — Wilderness
                                     •••• Non-wilderness
                                   Pacific Northwest (4B)
                                         Wilderness
                                         Non - wilderness
                                   Northern Rockies (4C)
                                    	Wilderness
                                    	Non-wilderness
                                   Central Rockies I4D)
                                    	Wilderness
                                    	Non - wilderness
                                   Southern Rockies (4E)
                                    	Wilderness
                                    	 Non-wilderness
                200
                         400
                                   600
7000
                                              California (4A)
                                               — Wilderness
                                               ••••  Non-wilderness
                                             Pacific Northwest (4B)
                                              	Wilderness
                                              	Non - wilderness
                                             Northern Rockies (4C)
                                              	Wilderness
                                              	 Non-wilderness
                                             Central Rockies (4D)
                                              	Wilderness
                                              	 Non-wilderness
                                             Southern Rockies (4E)
                                              — Wilderness
                                              	Non - wilderness
                                                                              200      400      600

                                                                                     Calcium (fjeg L~1)
                                                      800
                                                              1000
                                                                                          75

-------
Figure 5-11.   Inverse cumulative frequency distributions
             [1-F(x)] forsulfate (peg i"V /'" wilderness area
             (•	) and non-wilderness area (•	)
             lakes in California (4A). the Pacific Northwest
             (4B). the Northern Rockies (4C), the Central
             Rockies I4D). and the Southern Rockies (4E),
             Western Lake Survey - Phase I.
    1.0

    0.8 -

    0.6 -
 1-FM
    0.4 -

    0.2 -
    1.0

    0.8 -

    0.6 -
 1-FM
    0.4

    0.2 -

    1.0

    0.8 -

    0.6 -
 1-FM
    0.4 -

    0.2-

    1.0

    0.8 -
    0.6 -
 1-FM
    0.4
    0.2 -

    1.0

    0.8 -\
    0.6 -
 1-FM
    0.4 -

    0.2 -

    0.0
         California (4A>
          — Wilderness
          	Non - wilderness
        Pacific Northwest (4B)
          	Wilderness
          	Non-wilderness
        Northern Rockies (4C)
         	Wilderness
         	Non-wilderness
        Central Rockies (4D)

          	Wilderness
          	A/on - wilderness
        Southern Rockies (4E)
          	Wilderness
          	Non-wilderness
              —i—
               20
—i—
 40
—I—
 60
—i—
 80
                                              100
                     Sulfate (peq L'
             5.8   Sample Statistics for Special
             Interest Lakes

             Thirty-two special  interest lakes, selected on the
             basis of available  historical  data,  were  sampled
             during the WLS-I using protocols identical to those
             used for the  probability sample lakes.  Data from
             these lakes were used in the data validation process
             (Section 3.3).  Population estimates using data from
             these lakes were not calculated because the lakes
             were not selected as part of the probability sample.
             The data  presented here  are, therefore,  sample
             statistics (Table 5-23). The data for individual lakes
             are listed in Volume II.

             Special interest lakes were located at high elevations
             (M = 2903 m). As a group, these lakes markedly varied
             in watershed  area  (23 to 26571  ha) and lake area
             (1.0 to 5397 ha). The largest lake was Crater Lake,
             Oregon (ID number =  4B3-071), and  the largest
             watershed area was estimated for Fremont  Lake,
             Wyoming (4D3-076). Although the maximum site
             depth (109.7 m) was recorded for Fallen Leaf Lake,
             California  (4A3-047), site  depths for Crater  Lake,
             Oregon and Fremont Lake, Wyoming were too great
             to be measured with the sampling equipment.

             None  of  the  special  interest  lakes  were  acidic
             (minimum  ANC = 23.2 /ueq L"1). The  maximum ANC
             (1100/ueq L'1) was measured in Lake Catherine, Utah
             (4D3-079). All  special interest  lakes  had  pH
             exceeding  6.3. Sulfate was generally low for most
             lakes except Crater Lake, Oregon, where the value
             was 221.9 fjeq L~\  Most special interest lakes had
             low values for DOC, turbidity, and  true color, and
             consequently had high Secchi disk transparencies.
5.9  Statistics for Dilute Lakes

Throughout the West, many of the  lakes sampled
had very low ionic strength. Because such low ionic
strengths can  be indicative of low acid neutralizing
capacity, it  is  interesting to  examine the physical
and  chemical  characteristics of these  lakes.  A
subpopulation of "dilute" lakes was defined as those
lakes in  the target  population having conductance
< 10 fjS cm"1.  An estimated total of 2766 lakes  in
the West had conductance < 10 /uS crrf1 (Figures
5-13  to 5-17).  In contrast, the 20 percentile value
for conductance measured in the ELS-I was 20.7
/uS cm~1 (Linthurst et al. 1986). California contained
the  highest  percentage  of  dilute lakes (58.2%),
whereas only 7.2 percent of the lakes in the Southern
Rockies had conductance < 10 /uS cm"1.
                                                    Dilute lakes were more transparent than the target
                                                    population of lakes in each  subregion,  with the
                                                    exception of California (Table 5-24). Median Secchi
                        76

-------
Figure 5-12.   Concentrations (ueq L~1J of ANC, calcium, and sulfate in lakes in selected wilderness areas and national parks.
              Western Lake Survey -  Phase I. Numbers in parentheses after each area name are those shown in Figures
              2-2 and 2-3 and Table 2-4 in Section 2.2.3. First quintiles (Qt. 20th percentile). medians (M. 50th percentile).
              and fourth quintiles fQ4,  80th percentile), are given for each variable ( Q	1	D ).
                                                                       Qi   M  Qt
   California (4A)

Klamath Mountains (1 -3)
Southern Cascades (4-7}
Tahoe Area (8-9)
YosemiteArea(10-13)
Southern Sierra Nevada (.14-18) -
   Pacific Northwest (4B)

North Cascades Area (19-26)
Alpine Lakes (29)
Washington Cascades (30-34)
Oregon Cascades (35-42)

   Northern Rockies (4C)

 Northwest Montana (43-47)
 Bitterroot Area (48-51)
 Blue Mountains (52-53)
 Sawtooth Mountains (54)

  Central Rockies (4D)

 Southern Montana (55-56)
Northwest Wyoming (58-62)
Wind River Range (63-66)
Wasatch/Uinta Mtns. (67-68)

   Southern Rockies (4E)
Mount Zirkel (69)
Rocky Mountain Park (70-73)
Central Colorado (75-81)
S. Colorado/New Mexico (82-84) -
1 	 1 _T . _J 	
H-d
Ki
CH-Q



X-D


cH — n




0|Q £7

o .!..„_. . .__n
1 1 1 1 I
- 	 T 1 1
M 	 a
HH3
)-K3
n 1 n



— h- a
687

o-|-a

0 1 a 547
o+a




1 1 1
                              100  200  300 400

                                 ANCfueqL'')
  10   20  30  40   50

     Sulfate (ueq L'^)
100  200   300

Calcium foeq L''1)
disk transparencies for dilute lakes  in the  Pacific
Northwest, and in the Rockies were more than one
meter greater than for the target population (Section
5.6.2). Dissolved organic carbon and true color were
also low, factors which also contributed to the high
transparencies.

Acid neutralizing  capacity and Ca+z were lower in
the subpopulation of dilute lakes than in the total
population (Section 5.4.1 and 5.4.4), particularly in
the Northern and  Southern Rockies.  Median S04~2
concentrations in dilute  lakes   in  California, the
Pacific Northwest and the Northern Rockies were
6.2//eqL"1orless,  and were approximately 16//eqL~1
in the Central and  Southern Rockies. For dilute lakes
in the West as a whole, the CU value for SO,T2 was
13.0 peq L"1. Concentrations of extractable Al were
very low in dilute lakes in the West (M = 3.3 /JQ L"1).
Dilute lakes were associated with somewhat higher
elevations  than the target  population lakes as  a
whole. Dilute lakes in the Pacific Northwest occurred
at relatively lower elevations than in the other areas
(Table 5-24).  Median elevations for dilute lakes in
California and in the Central and Southern  Rockies
were greater than 2900 m. Watershed areas were
slightly smaller than those estimated for the entire
target population (Section 5.3), although lake areas
for the populations were comparable; consequently,
watershed  to lake area  ratios were  less  for dilute
lakes.
                                                                               77

-------
Table 5-23.    Sample Statistics for Physical and Chemical
              Variables for Special Interest Lakes8,
              Western Lake Survey - Phase I
Variable Units
Lake Elevation m
Watershed Area ha
Lake Area ha
Site Depth m
ANC neq L~1
PH
c\f*L —2 i —1
b(J* ueq L
i~ +2 i-1
La /yeq L
Al, ext. /^gL^1
DOC mg L"1
NOs" yueqL"1
NH4 jueq L ~
P, total pgL~1
True Color PCU
Turbidity NTU
Secchi disk
transparency m
Na+ f*eq L
K + /ueq L -1
Mg+2 /jeqL-1
Fe /jg L~1
Mn /ug L~1
Al, total /jg L~1
SiO2 mg L ~
DIG mg L~1
Cr /"eqL~1
F", total dissolved fjeq L~1
Conductance /jS cm"1
MINb
1424
23
1.0
2.4
23.2
6.36
1.4
10.5
0.0
0.24
0
0
0
0
0.1

1.6
5.6
1.5
4.9
0
0
3.5
0.16
0.46
0.8
0.2
3.4
Mc
2903
220
6.6
7.8
103.8
7.18
18.3
81.1
3.0
1.28
0.4
0
7.6
10
0.35

4.9
20.1
6.2
20.8
13.25
2.5
17.0
1.51
1.39
3.4
0.7
13.2
MAXd
3913
26571
5396.7
109.76
1100
8.73
221.9
675.8
9.4
4.28
21.4
7.1
41.4
25
5.85

15.2f
445.7
50.1
424.0
162.5
20
74.8
19.04
12.00
302.6
4.9
112.0
aSample size = 32.
bMIN = minimum (Oth percentile).
CM = median (50th percentile)
dMAX = maximum  (100th percentile).
e Depth was undefined for Crater Lake, Oregon (4B3-071) and
 Fremont Lake, Wyoming (4D3-076).
f Secchi disk transparency was undefined for Crater Lake,
 Oregon.

-------
Figure  5-13.    Locations of dilute lakes (conductance < lOftS cm'^ samples in California (4A), Western Lake Survey - Phase I.
               Geographic area covered by the target population is bounded by a dashed line.

-------
Figure 5-14.    Locations of dilute lakes (conductance < 10 /uS cm V sampled in the Pacific Northwest (4B). Western Lake
               Survey - Phase I. Geographic area covered by the target population is bounded by a dashed line.
                            80

-------
Figure 5-15.    Locations of dilute lakes (conductance < 10 (iS cm'*) sampled in the Northern Rockies (4C).  Western Lake
               Survey - Phase I. Geographic area covered by the target population is bounded by a dashed line.
                                                                                     81

-------
Figure  5-16.    Locations of dilute lakes (conductance < 10 /jS cm V sampled in the  Central Rockies (4D),  Western Lake
               Survey - Phase I. Geographic area covered by the target population is bounded by a dashed line.
                            82

-------
Figure 5-17.    Locations of dilute lakes (conductance < 10 /jS cm'*) sampled in  the Southern Rockies (4E), Western Lake
              Survey - Phase I. Geographic area covered by the target population is bounded by a dashed line.
                                                                  WY
                                          *-\\   .'    ''   /
                                   V         I    \
                                     -'UU
                                                                                 83

-------
Table 5-24.    Population Statistics for Chemical and Physical Characteristics of Dilute Lakes8, Western Lake Survey - Phase


California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West


California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West


California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
West


4A
4B
4C
4D
4E
4


4A
4B
4C
4D
4E
4


4A
4B
4C
4D
4E
4

N
1398
452
564
235
116
2766

N
1398
452
564
235
116
2766

N
1398
452
564
235
116
2766

QI
26.2
19.4
34.4
35.6
30.1
26.9

QI
1.0
2.8
1.9
0
0.2
1.0

QI
2.2
2.1
2.1
3.3
2.0
2.2
ANC (neq L~1)
M
45.1
33.0
46.6
52.3
45.8
45.8
Ext. Al (^g L~1)b
M
3.0
6.5
4.1
1.0
1.6
3.3
Lake Area (ha)
M
4.7
3.4
3.5
5.2
4.6
4.1
C.«0«qL-i) S04-*0-qL-')
Q4
61.9
58.8
68.2
58.6
60.9
63.7

Q4
5.1
11.4
8.6
2.6
3.9
6.9

Q4
13.3
10.9
6.3
12.4
8.0
11.8
QI
18.7
10.6
25.9
34.7
38.6
19.8

QI
0.58
0.33
0.75
0.50
0.49
0.54

QI
45
28
21
44
29
36
M
32.0
22.9
32.2
44.8
50.6
33.2
DOC (mg L"1)
M
0.85
0.97
1.15
0.73
0.66
0.92
Watershed Area
M
110
91
49
112
95
94
Q4
43.0
44.7
48.8
54.5
67.4
47.9

Q4
1.76
1.75
1.61
1.05
0.91
1.66
(ha)
Q4
337
217
101
476
114
247
QI
3.9
1.8
5.2
12.7
11.2
4.1
Secchi
QI
4.5
3.8
3.6
3.6
3.4
4.2

QI
2350
1276
1989
2919
3272
1828
M
5.2
5.4
6.2
16.2
16.0
6.8
Q4
9.0
15.8
8.1
23.1
18.8
13.0
Disk Transparency (m)
M
7.6
6.0
7.0
5.5
6.2
7.0
Elevation (m)
M
2999
1600
2197
3131
3531
2676
Q4
11.1
11.0
9.8
7.3
6.2
10.4

Q4
3295
1763
2599
3376
3761
3241
N  =  estimated number of dilute lakes.
QI =  first quintile (20th percentile).
M =  median (50th percentile).
Q4 =  fourth quintile (80th percentile).
aConductance < 10 ^S cm"1.
                             84

-------
                                          Section 6
                              Associations Among Variables
6.1   pH and Acid Neutralizing Capacity

6.1.1  Comparison of pH Measurements

For each lake sampled, pH was measured five times
on three  separate aliquots (Section 2.5). Because
pH is a non-conservative parameter, in situ mea-
surement  is preferred,  but  even under  optimal
conditions,  in  situ  measurements are difficult  to
perform with a high degree of accuracy. In situ
measurements were conducted by the helicopter
crews during the WLS-I, but portable pH meters were
not available to make in situ measurements on the
wilderness  lakes sampled by  the  ground crews
(Section 2.5.1.2). Thus, the surrogate measurement
for in situ pH  was the  "closed system" pH made
on the syringe aliquot that was unexposed to the
atmosphere (Section  2.5.1.2).  Results of the
comparison  of all  pH  measurements for  lakes
sampled by helicopter crews are shown in Table 6-1;
comparison between helicopter and ground sampled
lakes was discussed in Section 4.

Differences in the intercepts and slopes for regres-
sions of all pairs of measurements made in the field
and analytical laboratories were minor. As indicated
by the r2 values, the two open system measurements
of pH  made by the analytical  laboratories are
constant.  The regression for air-equilibrated pH and
closed system pH had a relatively low r2 value and
a slope of 1.00, indicating that, upon equilibration
with  the atmosphere, some lake samples gained CO2
while others lost COa-
All laboratory measurements showed relatively poor
agreement with the in situ  pH  measurements.
Although  in situ measurements generally  agreed
with closed  system  pH, the agreement among all
laboratory  measurements was lower  than  that
observed  in ELS-I  (Linthurst  et  al.  1986).  The
agreement among pH measurements (as indicated
by r2 values) was lower in the West compared to
that observed in the East and is probably attributable
to several factors. The range of pH measurements
in the West was much  less than that observed in
the East. Thus, the regression equations for the East
and West represent different populations of lakes.
A direct  comparison of pH measurements  with
eastern and western lakes should  be  restricted to
the same range. The pH values  commonly encoun-
tered  in  the West  (usually greater than 6)  also
showed greater variability among pH values in the
East. Additionally, the low conductance of western
lakes likely contributed to instability in the measure-
ment of pH (Brezinski 1983; Metcalf 1984).


6.1.2  Relationship Between pH and A cid
Neutralizing Capacity

At a given pH, ANC  is a linear function of the  sum
of carbonate species (Cj) where:

      CT = [C03~2] + [HC03~ ] + [C02] + [H2CO3]

If CT is fixed by holding the partial pressure of CO2
constant, a unique value of ANC is defined for each
Table 6-1.    Weighted Regression Statistics Comparing pH Measurements for Lakes Sampled Using Helicopters,
           Western Lake Survey - Phase I
pH Pair3
Dependent/Independent Variable
Closed system/acidity titration
Closed system/alkalinity titration
Closed system/air equilibrated
Closed system/in situ
Acidity titration/alkalinity titration
Acidity titration/air equilibrated
Acidity titration/in situ
Alkalinity titration/air equilibrated
Alkalinity titration/in situ
Air equilibrated/in situ
"n = 41 1, except for in situ pairs for which n = 406;
Intercept
-0.21
-0.19
-0.14
1.64
0.02
-0.14
2.60
-0.17
2.63
3.89
variables are defined in Volume I
Slope
1.053
1.050
1.004
0.773
0.996
0.981
0.619
0.986
0.615
0.481
II, Tables 2-1 and 5-1.
r2
0.866
0.862
0.671
0.798
0.994
0.820
0.653
0.828
0.644
0.467

                                              85

-------
pH. This condition can be achieved by equilibrating
the samples with a standard concentration of COa
in the atmosphere (referred to as air-equilibrated pH;
Section 2). The  comparison of air-equilibrated pH
to ANC (Figure 6-1) shows  that measurements of
these variables agree closely to the partial pressure
of C02 predicted on the basis of the standard air
used (10~35 atm). However, natural lake waters vary
widely in their partial pressure of COa resulting in
considerable scatter between closed system pH and
ANC (Figure 6-1). It is also likely that organic acids
contribute  to  deviations  from  the expected
relationship between air-equilibrated pH and ANC.
Most  of the  western lakes exhibit higher  partial
pressures of COa (i.e., oversaturation) than predicted
on the basis of COa in the atmosphere (1CT35 atm),
    as indicated by observations shown below the curve.
    The lakes that were undersaturated plot above the
    curve. The large variation in COa partial pressures
    for western lakes makes it difficult to generalize on
    patterns in the relationship between pH and ANC.


    6.2  Major Cations and Anions

    Although bicarbonate and Ca+2 are the principal ions
    in most western lakes, a more thorough understand-
    ing 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.
Figure 6-1.    Relationship between closed system pH and measured ANC for lakes with pH < 9.0 and ANC < 1000 fjeq i"1
            in all subregions. Western Lake Survey - Phase I. This same relationship is shown in the inset for air-equilibrated
            pH.
              8 -
           Hi
           2.
           to
              7 -
                                                                            600 800 7000
                                                                    ANC(ueqL^)
                -200
                                        200
    400

ANC (fieq L"1)
 I
600
                                                                            I
                                                                           800
1000
                      86

-------
6.2.1   Individual Example Lakes

The strength and focus of the Survey design is that
it  allows specific subpopulations  of lakes to be
described.  However, differences in lake chemistry
among subregions  also can be  evaluated by
examining the chemistry of individual lakes that are
"typical" of a population. Although no lake can be
representative  of other  lakes with respect to all
chemical attributes,  the  relative concentrations of
anions and cations observed among individual lakes
can be used to  highlight  similar features among
lakes.

Associations among  anions  and cations  can be
examined directly with ion bar charts. Ion bar charts
were developed for a single lake selected from each
of the  five  subregions (Figure 6-2). To standardize
the presentation, lakes were selected on the basis
of ANC and DOC values. Each lake selected had ANC
and DOC concentrations  approximately equal to the
first quintile  (20th  percentile) values for the
respective subregion (Section 5.5, Table  5-16). The
concentrations of cations and anions in each lake
are expressed as percentages of the total measured
ionic concentration,  which are shown adjacent to
the  bar chart  for each  lake.  Ions for which the
measured concentrations were generally less than
5  //eq  L"1  (Fe+3, Mn+2, Al+n,  NH4+,  N03~) are not
included in this presentation.

All lakes shown exhibit a small anion deficit (defined
as the sum of cations - sum of anions), presumably
attributable to the presence of unmeasured  organic
anions. When  organic anions are  estimated from
DOC and pH using the model by Oliver et al. (1 983),
much  of the deficit  is  explained. The absolute
magnitude  of this deficit is small (generally about
10 //eq L"1  in  these  lakes),  but because the lake
waters are so dilute,  particularly in California (4A)
and the Pacific Northwest (4B), the percentage of
the total anion concentration represented by organic
anions is moderate. An alternate explanation for the
anion deficit, i.e., analytical  error,  is discussed in
Section 6.2.7.

Bicarbonate  is  the  dominant  anion in the five
example lakes,  followed  by  SO4~2  and  then Cl~.
Despite a four-fold difference  in the concentrations
of total  ions among  these  example  lakes, the
proportions of anions are  similar, with the exception
of Champion Lake, Idaho (ID number =  4C2-047).
Unlike  the  other lakes shown here, in  which F~
concentrations are low,  this  lake exhibits  a high
concentration of F". Although F~ concentrations were
generally low in  western lakes (95.9% of western
lakes had  F~  <  10 //eq L"1),  moderately high
concentrations were observed  in the  Northern (4C)
and Southern Rockies (4E) where 7.2% and 17.1%,
Figure 6-2.   Ionic composition of one lake selected from
            each of the five western subregions surveyed
            during the Western Lake Survey - Phase I. Each
            lake was selected on the basis  of having
            concentrations of ANC (ueq i"V  and  DOC
            (mg L ~V approximately equal to the first quintile
            (Ct\, 20th percentile) values estimated for the
            subregional target population of lakes. Cations
            and anions are expressed as percentages of the
            total measured ionic equivalents (ueq L~^); ions
            with concentrations < 5 ueq i~1 are not shown.
            Champion Lake (4C2-047) is shown on the
            USGS topographic map as an unnamed lake
            and is listed as "No Name" in the data base
            and lake lists. Volume II.

                                       Total
                                       Ionic
                0  10  20  30  40 50  60 Equivalents

                                        (ueq Z."V
Swamp Lake, CA
(4 A1-037)
Airplane Lake, WA +
(4B1-053)
Champion Lake, ID +
(4C2-047)
Weyman Lake, UT +
f4D 1-046)
Tobacco Lake, CO  +
(4E3, 049)
                         67.7
                         108.3
                         144.5
                         181.8
                         280.2
    Cations (+)

    I Calcium
  0   JO  20  30  40  50  60
   Percent Ionic Equivalents


n           inn          nyi
j Magnesium Ba Potassium  Kj Sodium
    Anions (-)

    In Bicarbonate
                    Sullate
             | Chloride  E3 Fluoride
respectively, had F > 10 yueq L \ No lakes with F
> 10 fjeq  L"1 were sampled in  California, and F'
exceeded this concentration  in only one lake in the
Pacific Northwest.

Calcium comprises approximately one-half of the
total concentration of cations in  the lakes shown
here. In contrast  to  lakes in the East, Mg+2 is  a
relatively minor cation  in these lakes, except in
Weyman Lake (ID number = 4D1 -046), Utah in which
Mg+2 is the second most abundant cation. In the other
four lakes, Na+ is the  second  most abundant cation.
The excess  of  Na+ relative  to  Cl~ suggests that
                                                                            87

-------
watersheds, and  not marine  influences, are  the
major source of Na+ to these lakes. Concentrations
of K+  were moderately high  in  Tobacco  Lake,
Colorado (ID number = 4E3-049), exceeding even the
concentration of Mg+2.

6.2.2  Calcium and Magnesium

The relationship between Ca+2 and Mg+2 concentra-
tions in western lakes was highly variable (Figure
6-3). The regression coefficients (r , Table 6-2), were
similar to  those observed for the Northeast.  In all
subregions except the Northern Rockies, the lowest
r2 values are associated with the strata representing
the lowest alkalinity map class. This is consistent
with the  low  concentrations  of  Ca+2 and  Mg+2
observed in the lakes of these strata. The slopes for
Ca+2 and Mg+2 are also similar to  those observed
for the  Northeast and Upper Midwest, except for
California  where the slope approaches one.  How-
ever, the large differences in slopes and intercepts
apparent among strata within the same subregion,
illustrate the difficulty in developing a  model that
adequately describes a single relationship between
two variables applicable for the West as a whole.
6.2.3  Calcium and Sulfate

A weak positive relationship between Ca+2 and S04~2
was observed  in the West (Table 6-3), suggesting
that some  of  the  S04~2 is  derived from  mineral
sources. The sources of S04~2 could be weathering
of minerals such as sulfides or gypsum (Stumm and
Morgan 1981) or the deposition of wind-borne soils
containing gypsum. Gypsum  could also be a source
of Ca+2 in addition to that derived from carbonates.
Hidy and Young (1986) cite the high correlation
coefficient between Ca+2 and S04~2 in wet deposition
(r = 0.92) as evidence that most of the S04~2 is derived
from wind-borne soils. Because Ca+2andSO4~2occur
in equivalent amounts in gypsum (CaS04-2H20) and
anhydrite (CaS04), the slope  for this relationship in
precipitation should be  one. However, SO4~2 occurs
in greater amounts than Ca+2, casting some doubt
on  wind-borne gypsum as the  primary source of
S04~2 to some of these lakes (Oppenheimer et al.
1985). Nevertheless, S04~2 concentrations show  a
positive association with ANC (Figure 6-4) suggest-
ing  that watershed  sources of sulfate are important
to some lakes in the West.  In all subregions, the
highest median S04~2 concentrations are observed
 Figure 6-3.    Relationship between calcium   500
             (< 500 iieq L'1) and magnesium
             (< 250 peg L''1). Western Lake
             Survey - Phase I.
                                       400 -
                                       300 -
                                    
-------
Table 6-2. Regression Statistics for Calcium (< 500 /jeq L~1, Dependent) versus Magnesium
and Stratum, Western Lake Survey - Phase 1
Subregion/Stratum
California



Pacific Northwest



Northern Rockies



Central Rockies



Southern Rockies




4A1
4A2
4A3
4A
4B1
482
4B3
4B
4C1
4C2
4C3
4C
4D1
4D2
4D3
4D
4E1
4E2
4E3
4E
n
54
50
34
138
56
52
43
151
53
43
25
121
43
47
29
119
46
45
22
113
Intercept
53.0
33.3
36.6
51.0
40.4
55.7
89.4
48.4
21.4
53.8
91.5
68.3
62.5
14.7
52.9
35.7
49.5
75.5
55.4
56.8
SEa
12.9
8.7
19.3

12.2
16.2
31.6

5.4
12.0
33.2

9.0
16.4
25.9

11.1
21.8
25.4

Slope
1.870
1.335
1.462
1.424
1.585
1.370
1.675
1.897
3.277
2.313
1.333
1.557
0.773
2.673
1.975
2.088
1.761
1.705
2.057
2.004
(< 250 fjoq L"
SEa
0.706
0.282
0.292

0.529
0.338
0.415

0.397
0.291
0.350

0.317
0.421
0.345
0.167
0.355
0.333
0.344

1) by Subregion
r2
0.119
0.318
0.440
0.166
0.143
0.247
0.284
0.395
0.572
0.607
0.386
0.489
0.127
0.473
0.548
0.572
0.358
0.378
0.642
0.590
3SE = standard error.
Table 6-3.    Regression Statistics for Calcium (< 500 fjeq L 1, Dependent) versus Sulfate (< 100 jueq L 1) by Subregion and
             Stratum, Western Lake Survey - Phase I

Subregion/Stratum                       n        Intercept        SEa          Slope          SEa           r2

California                  4A1          54        38.8           5.6         1.339         0.091         0.807
                          4A2          50        38.3           9.8         3.781         1.108         0.195
                          4A3          37        77.4          16.6         3.333         0.698         0.395
                          4A          141        46.6                       1.345                       0.638

Pacific Northwest           4B1          57        47.4           9.8         1.529         0.416         0.197
                          4B2          52        58.7          15.8         2.067         0.520         0.240
                          4B3          43        77.3          17.0         3.016         0.323         0.680
                          4B          152        49.3                       2.884                       0.523

Northern Rockies           4C1          53        27.6           7.6         2.596         0.580         0.282
                          4C2          43        57.7          17.7         4.327         0.986         0.320
                          4C3          25       106.5          39.0         3.228         1.243         0.227
                          4C          121        68.6                       3.894                       0.305

Central Rockies             4D1          43        54.8          14.2         0.999         0.501         0.088
                          4D2          46        42.5          15.9         2.579         0.534         0.346
                          4D3          29       169.0          22.7         0.125         0.146         0.026
                          4D          118       106.1                       0.242                       0.050

Southern Rockies           4E1          46        25.9           9.5         2.289         0.268         0.624
                          4E2          45       120.0          20.9         1.113         0.362         0.180
                          4E3          22       134.7          22.5         1.023         0.280         0.400
	4E	113	118.2	1.104	0.347

aSE = standard  error.

-------
Figure 6-4.    Median sulfate concentrations
            (fjeq t"V for lakes  in three
            measured ANC classes for Cali-
            fornia (4A), the Pacific North-
            west (4B). the Northern Rock-
            ies (4C), the Central Rockies
            (4D), and the Southern Rockies
            (4E),  Western Lake Survey -
            Phase I.
50
40
                                           30
                                       I
                                       to

                                       I   20

                                       I
                                           10


-


-
1













^
§
1
'//////////////////////////////A



r
Acid Neutralizing
Capacity (fjeq L)
D - 50
H] > 50-200
g > 200


\





—
7T.
|
$
^«
^
I
Y//////////////////////////////A


—


—
/
«•
^
//
V
//
^/
^.*
/£
°v
in the lakes in the highest ANC class. Only  in the
Central  Rockies, the subregion  containing Yellow-
stone National Park, is there little difference in SC<4 '
concentrations among the ANC classes.  Additional
discussion  of SCV2 in the  Pacific  Northwest is
included in Section 7.2.

6.2.4  Acid Neutralizing Capacity Versus Base
Cations
In most surface waters, Ca+2 and Mg+z are the major
base cations and HCOs" is the dominant anion (Rodhe
1949). In many watersheds, Ca+2 and Mg+2 are the   6.2.5  Relative Abundance of Major Ions
        approaching a slope of one and a zero intercept (Table
        6-4). Lakes in the Central  and Southern  Rockies
        showed a small apparent deficit of ANC relative to
        the  sum of Ca+2 and Mg+2  However, examination
        of the standard errors from the strata for these two
        subregions  indicates that  only the  intercept for
        stratum  4E1  is  significantly less than zero. An
        adjustment  of the data for  possible  analytical bias
        may eliminate the apparent ANC deficit suggested
        in this analysis (Section 4.5.2; 4.6).
primary weathering products (Drever 1982), because
the weathering of calcite or dolomite results in the
production of  alkalinity in proportion to  Ca+2 and
Mg+2. Small amounts of calcite in watersheds may
determine weathering  products and  influence the
production of alkalinity (Drever and Hurcomb 1986).

The relationship between ANC and the sum of Ca+2
and  Mg+2 is plotted in Figures 6-5 to 6-9. This
regression has been used to infer acidification based
on departures from a near zero to negative intercept
and a slope reduced from one (Aimer et al. 1978).
This relationship  is indicative  of carbonic acid
weathering and also forms the basis of the predictor
nomograph of Henriksen (1980),  which  is  an
empirical model of acidification.

Lakes in the West generally showed a relationship
between  ANC and  the sum of  Ca+2 and  Mg+2
        The relative importance of major cations and anions
        among subregions can be examined by ranking the
        ions in order of decreasing concentration (Table 6-5).
        The relative abundances presented here are based
        on population estimates for the first quintile (20th
        percentile)  value  and the  median  concentration.
        Calcium is the dominant cation in all subregions at
        both the first quintile and median values. However,
        the relative abundance of Na+ and Mg+2 in California
        and the Pacific  Northwest differs from that  in the
        Central and Southern Rockies.

        The higher Na+  concentrations in the Pacific Coast
        subregions cannot be attributed solely to  marine
        contributions  of sea spray. Examining the relation-
        ship between lake CI" concentration and distance
        from  the  coast (Figure 6-10)  suggests  that the
        distance in the West over which sea spray influences
                     90

-------
Figure 6-5.    Relationship between mea-
              suredANC (< 200 ueq L~') and
              the sum of calcium and mag-
              nesium (ueq L~1) for lakes in
              California (4A). Western Lake
              Survey - Phase I. The alkalinity
              map class from which  each
              lake was selected is shown (O:
              < 100 ueq i"1; A.  100-199
              /jeq i~1;  X. 200-400 ueq L~\-
              Omernik  and Griffith 1986).
                                          200
                                          160 -
                                          120
                                        -j
                                        o-
                                           80
                                           40
Alkalinity Map
Class (ueq L'^)
D < 100
A 100 - 199
X 200- 400
                                                           40          80          120          160

                                                                  Calcium + Magnesium (ueq L~^)
                                                                                                           200
200
160 -
120-
 80
 40-
            Alkalinity Map
            Class (ueq L'^)
           D <100
           A 100 - 199
           X 200-400
              i      i      I      l      I      I      I
                   40          80          120
                        Calcium + Magnesium (ueq L'^)
                    Figure 6-6.   Relationship between mea-
                                  sured ANC (< 200 ueq L"1) and
                                  the sum of calcium and mag-
                                  nesium (ueq L' V for lakes in the
                                  Pacific  Northwest  (4B),
                                  Western Lake Survey - Phase /.
                                  The alkalinity map class from
                                  which  each lake  was selected
                                  is shown fO: <  100 ueq L'1;
                                  A. 100-199 ueq L~\- X. 200-
                                  400 ueq  L~\- Omernik and
                                  Griffith 1986).
                                                     160
               200
                                                                                    91

-------
Figure 6-7.    Relationship between mea-
              sured ANC (< 200 ueq O and
              the sum of calcium and mag-
              nesium (ueqL~1) forlakesin the
              Northern Rockies (4C), West-
              ern Lake Survey - Phase I. The
              alkalinity map  class from
              which each lake was selected
              is shown fC\: < 100 ueq L'\-
              A.  100-199 ueq L'\- X: 200-
              400 fjeg  L'\- Omernik and
              Griffith 1986).
                                              200
                                 160 -
                                              120-
                                           -j

                                           I
                                               40-
A Ika Unity Map
Class (ueq L'^J
n <'oo
A 700 - 199
X 200-400
                                                              40
                                                               i
                                                             80
                                                                                      120
                                                                                                   160
                                                                                                              200
                                                                    Calcium + Magnesium (ueq L'1)
   200
   160 -
   120 -
 t>
I
    80
     40
Alkalinity Map
Class (ueq L''1)
D <700
A 700-733
X 200-400
              Figure 6-8.   Relationship between mea-
                            sured ANC (< 200 ueq L~*) and
                            the sum of calcium and mag-
                            nesium (ueq L' V for lakes in the
                            Central Rockies (4D). Western
                            Lake Survey - Phase I.  The
                            alkalinity  map class from
                            which  each lake was selected
                            is shown (O: < 700 ueq i~V
                            A: 100-199 ueq i"1; X: 200-
                            400 ueq  L~\'  Omernik  and
                            Griffith 1986).
                     I      1      I      f      I      l
                    40          80          120

                         Calcium + Magnesium (ueq L''1)

                         92
                                           160
                                                       200

-------
Figure 6-9.    Relationship  between mea-    200
               suredANC (< 200 fjeq L~^) and
               the sum of calcium and mag-
               nesium (fieq L ' V for lakes in the
               Southern Rockies (4E). West-
               ern Lake Survey - Phase I. The
               alkalinity map  class from     160 -
               which each lake was selected
               is shown (O:  < 100 iieq L~\-
               A.- 100-199  iieq L~\- X: 200-
               400  neq I"1;  Omernik and
               Griffith 1986).
                                              120 -
                                           t>
                                           1
                                              80
                                               40
Alkalinity Map
Class (ueqL'1)
D <100
A 100 - 199
X 200 - 400
                                                               40          80          120

                                                                     Calcium + Magnesium (ueqL'^)
                                             160
200
Table 6-4.    Regression Statistics for ANC (< 200 fjeq L~', Dependent) versus the Sum of Calcium and Magnesium (< 200
              //eq L~1) by Subregion and Stratum, Western Lake Survey - Phase I
Subregion/Stratum
California



Pacific Northwest



Northern Rockies



Central Rockies



Southern Rockies




4A1
4A2
4A3
4A
4B1
4B2
4B3
4B
4C1
4C2
4C3
4C
4D1
4D2
4D3
4D
4E1
4E2
4E3
4E
n
49
44
20
113
53
41
16
110
52
34
12
98
41
38
13
92
40
26
10
76
Intercept
8.7
7.1
0.9
8.0
6.2
2.0
-2.7
3.8
9.3
11.3
-6.4
4.5
-3.2
-17.8
3.7
-9.9
-22.9
21.5
6.8
-11.4
SEa
4.5
4.2
4.0

5.2
9.0
19.1

3.4
6.1
9.1

9.3
10.6
15.2

6.4
12.8
18.3

Slope
0.968
1.131
1.339
1.020
0.978
0.997
1.054
0.995
0.943
1.046
1.109
1.046
0.858
0.993
0.870
0.935
1.089
1.102
0.869
1.007
SEa
0.068
0.060
0.059

0.058
0.088
0.150

0.047
0.068
0.090

0.089
0.082
0.127

0.069
0.108
0.147

r2
0.813
0.895
0.966
0.827
0.847
0.769
0.780
0.815
0.888
0.881
0.939
0.905
0.705
0.804
0.811
0.790
0.867
0.812
0.814
0.835
aSE = standard error.
                                                                                       93

-------
Table 6-5.   Relative Abundance of Major Cations and Anions  Within  Subregions Based on Population Estimates of
           Concentrations at the First Quintile (Qi) and Median (M) Values, Western Lake Survey - Phase I
                                         Cations8
                                                    Anions8
Subregion
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies

4D
4B
4C
4D
4E
Qi
Ca > Na > Mg ~ K
Ca > Na > Mg > K
Ca > Na ~ Mg > K
Ca > Mg > Na > K
Ca > Mg > Na > K
M
Ca>Na>Mg~K
Ca >Na>Mg>K
Ca>Mg>Na>K
Ca>Mg>Na>K
Ca >Mg >Na> K
Qi
HC03>S04~CI
HC03>CI ~SC>4
HC03>SO4~CI
HC03>SO4>CI
HCO3>S04>CI
M
HCO3>SO4~CI
HCO3>SO4~CI
HC03>S04>CI
HCO3>SO4>CI
HCO3>S04>CI
West
Ca > Na ~ Mg > K
Ca>Mg ~Na >K
HC03 > SC>4 ~ Cl
HCO3>S04>CI
'—indicates difference between estimates is <5/jeq L
lake chemistry is  minimal beyond  50 km  inland.
Brakke and  Waddell (1985) found that beyond 50
km inland  from  Puget Sound, lake water Cl"
concentrations in the Pacific Northwest showed no
further decline with  distance. The terrestrial
contribution of Na+to lakes in the Pacific Northwest
is evident from the plot of Na+ versus distance from
the coast (Figure 6-11). Sodium concentrations in
some lakes  over 70 km from the coast exceed 50
//eq L"1. The source of Na+ in some western  lakes
not influenced by sea spray has been  attributed to
weathering  products from  plagioclase  (Garrels and
Mackenzie 1967; Nelson 1985). Further evidence for
mineral sources of sodium to some lakes is presented
in the relationship between Na* and Cl" in the Pacific
Northwest.  The  observed  ratio departs from  that
expected on the basis of the ratio of these ions in
seawater (Figure 6-12).

In the Northern Rockies, Na+ is more abundant than
Mg+2 at the first quintile,  which is also consistent
with weathering of plagioclase minerals. Most of the
low ionic strength lakes in the Northern Rockies are
located over  the   Idaho  Batholith,  the  granitic
composition of which  is  similar  to the  batholith
underlying the Sierra Nevada (Hunt 1974). For lakes
with high total ionic concentration, Mg+2 is  more
abundant than Na+ because the lakes  appear to be
located  in areas underlain, in  part, by ultramafic
rocks (Hunt 1974).

The relative  abundance of major anions is the same
in all subregions except in the Pacific Northwest,
where Cl"  was more  abundant than  SO4'2. In
Northern Washington, the  dominant source of Cl"
is sea salt, whereas  in   the  Oregon  Cascades
relatively high amounts of Cl" are derived  from
weathering  (Nelson 1985). The  uniform relative
abundance of major anions and cations observed in
the West differs considerably from that observed in
the East where marked  differences  within,  and
among,  regions were  observed  (Linthurst  et al.
1986).
                       6.2.6  Relationships Among Major ions

                       The ratios of major anions and cations are presented
                       as trilinear plots (ternary diagrams) in Figures 6-13
                       to 6-17.  Like the ion bar charts (Figure  6-2), these
                       plots include only those ions for which the measured
                       concentrations were typically 5 /ueq L"1 or greater.

                       The axes on  the  plots represent the ion or pair of
                       ions expressed as percent of total ionic  concentra-
                       tion. The values increase  in the direction of  the
                       arrows from  0 to 100 percent. Points representing
                       the percent ionic composition are plotted for cations
                       on  one triangle, for anions on the second triangle,
                       and for  the  pairs of cations and  anions on  the
                       parallelogram. Thus, a lake for which all  of the total
                       ionic concentration is due to dissolved CaCOa would
                       appear in  the extreme lower left  portion of each
                       triangle and would appear in the left corner of  the
                       parallelogram. A lake acidified by sulfuric acid would
                       be  expected to appear near  the apex of  the
                       parallelogram. Although these trilinear plots do  not
                       show differences in concentrations or include other
                       ions of  possible  importance, they  are  useful  for
                       displaying patterns among the major ions.

                       The ion  ratios were relatively uniform  among  the
                       five western subregions. In nearly all cases, the lakes
                       can be described as calcium bicarbonate systems
                       (i.e., Ca+2  and HCOs" comprise over 50% of  the
                       cations and anions, respectively). Exceptions to this
                       pattern, indicated as circled points on the plots, were
                       noted for each subregion and are described below.
                       These outliers were subjectively identified  based on
                       separation from the predominant pattern for a given
                       subregion.

                       The selected  outliers for lakes in California display
                       an  interesting geographic split: the outlier lakes on
                       the cation  plot were located in either the California
                       Cascades (4A3-011 and 4A3-013 in Lassen National
                       Park,  and 4A2-005  and  4A2-060 in  Caribou
                       Wilderness Area) or in the Klamath Mountains (4A3-
                      94

-------
 'ۥ?
«i|l
•S o ^ *
*i^
sis-

              ,<0
            ll.ll!
            lilt!
                              l»  1
                 \   \   \
                                         8
                                        ^§
-8!
   i
I

40

nc
                                        _ o
 8 ••1
EteSl
O J *4

"••S€*
t8|^
•* .g < Q.
a o "5 o '«



          1
          
-------
Figure 6-11.
Concentration of sodium
f/jeq L~^)  versus distance
(<  100 km) from Puget
Sound or the Pacific Ocean
(whichever is less) for lakes
in five geomorphic units in
the  state of Washington in
the  Pacific Northwest (4B).
Western  Lake  Survey -
Phase I. Values are shown for
the  Wenatchee Mountains
(9). the Middle Washington
Cascades ( &.) Cascades( +),
the Puget Lowlands fX), and
the Olympic Mountains (O).
Distance from the coast was
determined from a SAS file
(SAS Institute 1985).
   100


   90


   80


   70



   60


^ 50 ~

 b-
 I: 40 -

 |

 I 30 ~


   20 -


   10
                                                             Pacific Northwest (4B)
                                                             B Wenatchee Mountains
                                                             A Middle Washington Cascades
                                                             + North Washington Cascades
                                                             X Puget Lowlands
                                                             Q Olympic Mountains
                                                                \i  "
                                                                 •.   a  •
 )
10
 ]
20
 ]
30
 \
40
                                                                       SO
                                                                            60
                                              \
                                             70
                                                                          \
                                                                          80
                                                                                             90   100
                                                      Distance from Coast (km)
007, 4A3-058 and 4A3-060). These lakes contain
a  relatively  high proportion of  Mg+2, which is
abundant in the ultramafic  rocks found in  these
areas. The two outliers on the anion plot, attributable
to a high proportion of SO4~2, were both found near
Yosemite National Park. Kidney Lake (4A1-014) had
a high concentration of SO4~2 (93 jueq L"1). Lake 4A1 -
005 (No Name) was extremely dilute (conductance
= 4.0 pS cm"1) and, with  a SC>4~2 concentration of
only 10 fjeq L~1, SO4~2 was the dominant anion.

The outlier  lakes  in the Pacific Northwest  had
different  chemistries than  the  outlier lakes in
California. Trout Lake (4B1 -031) has high ANC (579
/jeq L"1) and is located in the Middle  Washington
Cascades. The high magnesium concentration (401
,ueq L"1) was attributable to the watershed bedrock
containing dunite. The high [Na+ + K+] ratios in five
lakes located in  the  Oregon  Cascades  cannot be
explained by sea spray. Four of these lakes  (4B1-
004, 4B1-005, 4B2-004,  and 4B2-005), located in
or near the Diamond Peak Wilderness Area, were
very dilute (conductance = 2.2 to 4.5 //S  cm"1) with
only moderate concentrations of Na+ (10 to  20
yueq L"1). Only Hosmer  Lake (4B1-011) located on
the eastern  slope  of the Oregon  Cascades  near
Waldo Lake Wilderness Area, contained high  Na+
(219/jeqL"1).
                                      The Northern Rockies contained three outliers, all
                                      located in the Selway-Bitterroot Wilderness Area.
                                      The three lakes were dilute (conductance = 5.5 to
                                      9.5 fjS cm"1)  and  had relatively low proportions of
                                      Na+ (Jeanette Lake, 4C1 -015) or SO4"2(Holloway and
                                      Little Carlton  Lakes, 4C1 -010 and 4C1 -011).

                                      Four lakes in the  Central  Rockies had atypical ion
                                      ratios. Three of these (4D3-016, 4D3-017, and 4D3-
                                      052) are located in Yellowstone National Park. Lake
                                      4D1-013 was a  dilute lake (conductance = 7.2
                                      (iS cm"1)  located in the Bridger Wilderness Area in
                                      the Wind River Range. The three lakes in Yellowstone
                                      identified as outliers on the trilinear plot were very
                                      different from one  another. Fern Lake (4D3-017) was
                                      the only  acidic  lake  (ANC = -24 ^eq  L"1)  sampled
                                      in the West. It was affected by hot springs, and the
                                      S04~2 concentration was 818 yueq L~1. Nymph Lake
                                      (4D3-016) also had a high S04~2concentration (2909
                                      jueq L"1),  but  had an ANC value of 1341  /jeq L~\
                                      In  contrast, Goose Lake (4D3-052) was a sodium
                                      bicarbonate lake with low S04 2(30^eq L"1).

                                      Several lakes in the Southern Rockies were outliers
                                      on the anion  plot.  Black Lake #2 (4E3-022), located
                                      south of the Eagles Nest Wilderness Area,  was very
                                      alkaline (pH  =  9.13, ANC  =  682 yueq  L"1).  The
                                      remaining outlier  lakes, located in the San  Juan
                      96

-------
<0

I
O
OC
1
I
i
 -
« S!
€.5
                         1

                                                                                     10
                                                                                     <*
                                                                                     10
                                                                                     10
                                                                                     t\l
                                                     O-
                                                     01


                                                     I
                                                                                        g
                       8
                             o
                             O)
o
03
o    Q
10    >

1 barl) uinipos
3.
0-8
«
45*

II
•c .u
w-5
O K-
o ^

yjl
                         I


                         1
                         -c


                         I
•
                                                                                     «0
                                                                                     CM
                                                                                        I
                                                                                   _ o
to
s

I
                                  o    o
                                  00    IS
           o    o
           
-------
Figure  6-13.    Trilinear diagram showing the
               relative abundance of major
               anions and cations for lakes
               in California (4A),  Western
               Lake Survey - Phase I. Ratios
               are expressed as  percent of
               total ionic  concentration,
               increasing from  0 to  100
               percent along the axes in the
               direction of the arrows. Lakes
               for which the total ionic com-
               position cannot be described
               as being dominated by cal-
               cium bicarbonate are circled
               and their lake ID numbers are
               given.
                                 4A2-005
                                 4A3-060
                                             4A3-011
                                             4A3-013
                                            4A3-058—I
                                                                                                   4A1-014
                                                                                                         4A1 -005
                             //\\
4B'1-031
                                                             4B'1-005
                                                                         Figure 6-14.     Trilinear diagram showing the
                                                                                         relative abundance of major
                                                                                         anions and cations for lakes
                                                                                         in the Pacific Northwest (4B).
                                                                                         Western   Lake  Survey  -
                                                                                         Phase I. Ratios are expressed
                                                                                         as percent of total ionic con-
                                                                                         centration, increasing from 0
                                                                                         to 100 percent along the axes
                                                                                         in the direction of the arrows.
                                                                                         Lakes for which the total ionic
                                                                                         composition  cannot be des-
                                                                                         cribed as being dominated by
                                                                                         calcium  bicarbonate  are
                                                                                         circled and  their  lake  ID
                                                                                         numbers are given.
                           \4B1-004
                            4B2-004
                           \ 4B2-005
                            4B1-011
                        98

-------
Figure 6-15.    Trilinear diagram showing the
               relative abundance of major
               anions and cations for lakes
               in the Northern Rockies (4C).
               Western Lake Survey  -
               Phase I. Ratios are expressed
               as percent of total ionic con-
               centration, increasing from  0
               to 100 percent along the axes
               in the direction of the arrows.
               Lakes for which the total ionic
               composition cannot be  de-
               scribed as being dominated by
               calcium  bicarbonate  are
               circled and their  lake ID
               numbers are given.
//\\
                                  4C1-010
                                  4C1-011
                                                                     4C1-01S
                                                                        Figure 6-16.
                             //\\
                                                                           4D3-046
                                       4D3-052
                                '403-016
                              4D3-017

                          ' 4D3-OS2
                  Trilinear diagram showing the
                  relative  abundance of major
                  anions and cations for lakes
                  in  the Central Rockies (4D).
                  Western Lake Survey -
                  Phase I.  Ratios are expressed
                  as percent of total ionic con-
                  centration, increasing from 0
                  to 100 percent along the axes
                  in the direction of the arrows.
                  Lakes for which the total ionic
                  composition cannot be  de-
                  scribed as being dominated by
                  calcium  bicarbonate  are
                  circled  and  their  lake  ID
                  numbers are given.
                                                                                     99

-------
          453-053

    4E2-040

       4E2-059
                                                               Figure 6-17.
                                                               4E3-053
                                                               4E3-040
                                                                4E2-059
                                                                 4E1-040
                                                                 4E1-026
                                                                  4E1-022
                         Tfilinear diagram showing the
                         relative abundance of major
                         anions and cations for lakes
                         in the Southern Rockies (4E).
                         Western Lake Survey -
                         Phase I. Ratios are expressed
                         as percent of total ionic con-
                         centration, increasing from 0
                         to 100 percent along the axes
                         in the direction of the arrows.
                         Lakes for which the total ionic
                         composition  cannot be de-
                         scribed as being dominated by
                         calcium  bicarbonate  are
                         circled and  their lake ID
                         numbers are given.
                 Ca
                                             Cl
Mountains,  had  high relative concentrations  of
sulfate. With the exception of Columbine Lake, (4E2-
040, ANC = 34 /ueq L"1) these lakes had relatively
high ANC (> 138 //eg L"1). The high S04~2 concen-
trations may have been derived from the Tertiary
volcanic  bedrock  or sulfide deposits in the area
(Section 7.5).

6.2.7  Dissolved Organic Carbon

6.2.7.1   Dissolved  Organic Carbon  and True
Color—Water passing through  plant materials
leaches  complex organic  compounds that are
resistant  to  decomposition.  These  organic  com-
pounds often contain functional groups that  impart
a brown  color to water (Gjessing 1976).  Because
of this property,  color has often been  used as a
surrogate for dissolved organic carbon (DOC).

Regression equations  were used to compare the
relationships between dissolved organic carbon and
true color for lakes in  each stratum and subregion
(Table  6-6). As expected, the variance explained  by
the regression  equations  is low  due to the small
range  of  DOC and color  in western lakes (Figure
6-18).

6.2.7.2   Anton Deficit—The  principle  of electro-
neutrality states that the sum of cation equivalents
should equal the sum of anion equivalents, assuming
all  contributing  ions  are  measured accurately.  In
addition to serving as an analytical tool for quality
assurance checks (Sections 3 and 4), evaluation of
ion balances provides additional insight into regional
lake chemistry.

There is a close agreement between the sum of
anions and the sum of cations among all subregions
of the West (Figure 6-19) with a minimum r2 of 0.983
in the  Southern  Rockies  (Table 6-7).  In  most
subregions the intercept was different from zero and
the slope was significantly  less than one. Although
attenuation in the slope, caused by errors in the X
axis, would contribute to a small  reduction in the
slope,  the  magnitude of the  deviation  from  1.0
suggests that the anion deficit may  be related to
factors other than organic anions.

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 bias,
anion deficit  reflects the presence of unmeasured
organic anions.  Table 6-8  shows population  esti-
mates of the first  and fourth quintiles (Qi and Q4)
and the median for anion  deficit by subregion.  As
expected, due to the low color and high transparency
in western lakes, the magnitude of the anion deficit
was  very  low.  The  lowest anion deficits  were
observed  for  California,  whereas  the  Southern
Rockies showed the  greatest  anion deficit. The
magnitude of the  anion deficit among subregions
was roughly proportional to the total  ionic concen-
                      100

-------
Table 6-6.
Regression Statistics for DOC (< 10 mg L'\ Dependent) versus True Color (< 50 PCU) by Subregion and Stratum,
Western Lake Survey - Phase I
Subregion/Stratum
California



Pacific Northwest



Northern Rockies



Central Rockies



Southern Rockies




4A1
4A2
4A3
4A
4B1
4B2
4B3
4B
4C1
4C2
4C3
4C
4D1
4D2
4D3
4D
4E1
4E2
4E3
4E
n
54
51
41
146
58
53
46
157
53
50
38
141
43
47
35
125
46
51
40
137
Intercept
0.8
1.5
1.6
1.0
1.2
1.1
1.0
0.9
0.1
0.2
0.2
0.119
0.2
0.4
0.5
0.3
0.2
0.3
0.2
0.2
SEa
0.1
0.2
0.2

0.2
0.2
0.3

0.2
0.3
0.5

0.2
0.2
0.3

0.1
0.2
0.4

Slope
0.143
0.193
0.062
0.148
-0.021
0.091
0.120
0.103
0.125
0.147
0.192
0.181
0.123
0.141
0.120
0.138
0.103
0.124
0.158
0.152
SEa
0.018
0.022
0.017

0.031
0.020
0.021

0.016
0.029
0.042

0.022
0.016
0.020

0.010
0.019
0.023

r2
0.558
0.611
0.250
0.506
0.008
0.290
0.406
0.290
0.552
0.345
0.362
0.360
0.432
0.618
0.593
0.613
0.698
0.465
0.559
0.560
aSE = standard error.
Figure 6-18.    Relationship between DOC
               (< 10 mg L~') and true color
               (< 50  PCU) for lakes in all
               subregions. Western Lake
               Survey - Phase I.
                                                    10
                                                     8~
                                                  '~j 5
                                                     2~
                                                                           •  i
                                                                                     -
                                                                  10
20         30

Color (PCU)
                                                                                      1
                                                                                     40
                                                                                                            50
                                                                                      101

-------
Figure  6-19.    Relationship between the sum of anions (< ZOO peg i~V and the sum of cations (< 200 fjeq L ^ for lakes
               in California ( • ). the Pacific Northwest ( A ). the Northern Rockies ( + ),  the Central Rockies ( X ). and
               the Southern Rockies fO),  Western Lake Survey - Phase I. Points falling below the 1:1 (diagonal) line indicate
               an anion deficit, and above the line, a cation deficit. Inset shows the same relationship for anions ^ 2000
               fjeq i'1 and cations < 2000 peq L~\
              200
         -j
         o-
         <
         o
              180 _
              160 _
              140
              120 _
              100 _
               80 _
               60 _
               40 -]
               20 _
                     2000
                     1000
                                 •  California (4A)
                                 A  Pacific Northwest (4B)
                                 4-  Northern Rockies (4C)
                                 ^  Central Rockies (4DI
                                 E  Southern Rockies I4E)
                             20
\

40
\

60
 80

Sum of Anions (fjeq i"
I         I        I
100      120     140
                                                                                         160
                                            180      200
                         102

-------
Table 6-7.    Regression Statistics for the Sum of Major Anions (< 2000 /Jeq L~1, Dependent) versus the Sum of Major Cations'
            (< 2000 A»q L~1) by Subregion and Stratum, Western Lake Survey - Phase I
Subregion/Stratum
California



Pacific Northwest



Northern Rockies



Central Rockies



Southern Rockies




4A1
4A2
4A3
4A
4B1
4B2
4B3
4B
4C1
4C2
4C3
4C
4D1
4D2
4D3
4D
4E1
4E2
4E3
4E
n
54
53
41
148
59
53
45
157
53
50
38
141
43
46
33
122
46
51
35
132
Intercept
-9.4
-9.3
-8.0
-5.0
-13.6
-4.6
3.4
-3.4
-2.2
-6.5
-2.9
-3.8
8.5
-18.1
8.2
7.7
-5.8
0.6
-8.9
-7.34
SEb
2.5
2.2
4.3

2.6
2.7
6.0

1.8
3.2
14.4

6.3
7.1
18.5

5.5
11.7
22.4

Slope
0.989
0.912
0.908
0.922
0.982
0.885
0.882
0.895
0.879
0.914
0.902
0.904
0.848
0.964
0.828
0.832
0.925
0.872
0.906
0.904
SEb
0.017
0.009
0.008

0.013
0.011
0.011

0.018
0.008
0.017

0.045
0.036
0.028

0.033
0.030
0.028

r2
0.985
0.995
0.997
0.992
0.990
0.992
0.993
0.993
0.980
0.996
0.988
0.991
0.897
0.942
0.966
0.972
0.946
0.947
0.970
0.973
aOnly anion and cations with individual concentrations > 5 j*eq L 1  are included in the sums.
bSE =  standard error; not calculated for subregions (Section 5.1.2.1).
Table 6-8.   Population Statistics for Anion Deficit,
           Western Lake Survey - Phase I
                         Anion Deficit (^eq L""1)
Subregion
M
California 4A
Pacific Northwest 4B
Northern Rockies 4C
Central Rockies 4D
Southern Rockies 4E
4.5
7.0
9.6
5.4
10.2
10.2
22.0
32.4
19.6
38.6
22.2
40.5
82.4
38.7
132.3
West
                     6.9
                                19.6
                                             51.1
Q, = first quintile (20th percentile).
 M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
tration. This observation  is  consistent  with the
possibility that some of the  anion  deficit  can be
attributed to analytical error (Sections 4.5 and 4.6).

It can be inferred that the anion deficit results from
the presence  of  unmeasured  organic anions if a
relationship between anion deficit and DOC can be
established. To  evaluate further  the  presence of
unmeasured anions among subregions, anion deficit
was plotted and  regressed on  DOC (Figure  6-20,
Table 6-9). The  regression results indicate a poor
relationship between DOC and anion deficit in the
West. This poor fit can be attributed, in part, to the
low concentrations of DOC and total ions observed
in the West. However, it also suggests that the anion
deficit in the West may be attributable to an analytical
error. Substituting ANC for HCOa , computed using
measured  DIG, in the computation  of the  sum  of
anions reduces the anion deficit by approximately
70 percent. Other studies of low ANC  lakes in the
West report close agreement (i.e., approximately 5%
difference) between the sums of anions and cations
(Baron 1983; Brakke and Loranger 1986; Nelson and
Delwiche 1983; Duncan 1985; Meglen et al. 1985;
Melack et  al. 1985).  The results  of these  studies
further  support the contention that  the relatively
small anion deficit observed here is an artifact  of
a small analytical error in the  measurement of a
major cation (e.g., Ca+2) or anion (e.g., HC03" which
is computed from DIG).

6.3   Geology

Early investigations  of the  sensitivity of  aquatic
resources to acidic deposition implied a relationship
between the composition of bedrock and the ANC
of surface water (Galloway and Cowling 1978). This
relationship  can  be  affected by surficial  geology
(Shilts 1981),  hydrologic flow paths  (Chen et al.
1984), soil type (McFee 1980), or vegetation (Krug
and Frink 1983). These factors contribute to the lack
of a good relationship between bedrock geology and
water chemistry  in eastern North America. These
factors would be  expected to exert less influence
on lakes in the mountainous western United States,
because soil development is generally poor, glacial
till  is  often  absent,  and bedrock exposures are
common.
                                                                            103

-------
Figure  6-20.    Relationship  between  DOC
               (mg L~^) and anion deficit    10
               l/jeq L'^j for lakes in all sub-
               regions with  total measured
               ionic  equivalents  <  500     9
               fjeq  L~\   Western   Lake
               Survey -Phase I. Anion deficit
               is the difference (fjeq i"V     8
               between the sum  of  mea-
               sured anions and the sum of
               measured cations.               7
                                          I
                                              3 ~
                                              2 -
                                              1  -
        t
        .%'•
        I .
v **.•:**•*"
»*    *
                                                                                .
  10
        20
I
40
l
50
                                                                    30

                                                                    Anion Deficit {peq L''1
i
60
                                       70
80    90    100
Table 6-9. Regression Statistics for DOC (< 10 mg L~\ Dependent) versus Anion Deficit !
and Stratum, Western Lake Survey - Phase I
Subregion/Stratum
California



Pacific Northwest



Northern Rockies



Central Rockies



Southern Rockies




4A1
4A2
4A3
4A
4B1
4B2
4B3
4B
4C1
4C2
4C3
4C
4D1
4D2
4D3
4D
4E1
4E2
4E3
4E
n
54
51
37
142
58
53
44
155
53
47
31
131
43
47
27
117
46
48
26
120
Intercept
0.5
1.4
1.6
0.7
0.6
1.3
1.9
1.1
0.6
1.1
1.6
1.3
0.8
1.8
1.9
1.7
1.0
1.2
2.0
1.9
SEa
0.1
0.4
0.3

0.2
0.3
0.5

0.2
0.2
0.6

0.2
0.4
0.3

0.2
0.2
0.2

Slope
0.052
0.042
0.008
0.044
0.031
0.015
0.008
0.021
0.044
0.013
0.013
0.017
0.029
0.019
0.001
0.002
0.023
0.007
-0.009
-0.008
[< 100 yueq L'1
SEa
0.009
0.014
0.007

0.008
0.008
0.011

0.010
0.006
0.013

0.009
0.012
0.002

0.008
0.006
0.003

) by Subregion
r2
0.416
0.164
0.039
0.298
0.191
0.060
0.011
0.099
0.267
0.104
0.031
0.063
0.193
0.058
0.012
0.005
0.158
0.032
0.210
0.147
°SE = standard error; not calculated for subregions (Section 5.1.2.1)
                        104

-------
The  relationship between bedrock  geology and
lakewater chemistry was evaluated by comparing
WLS-I lake chemistry to bedrock class (Norton et al.
1982). The bedrock classification scheme used by
Norton et al. (1982) is as follows:

   Class  1—Low to no  acid  neutralizing capacity
            (granitic gneiss, quartz sandstones or
            equivalent).

   Class  2—Medium to low acid neutralizing capac-
            ity (sandstones, shales, conglomerates,
            or their metamorphic equivalents, high-
            grade metamorphic felsic to interme-
            diate volcanic  rocks, intermediate
            igneous rocks, calc-silicate gneisses
            with no free carbonate phases).

   Class  3—High to  medium  acid  neutralizing
            capacity (slightly calcareous rocks, low-
            grade intermediate to mafic volcanic
            rocks, ultramafic  rocks, glassy volcanic
            rocks).

   Class  4—"Infinite" acid  neutralizing  capacity
            (highly fossiliferous sediments or meta-
            morphic  equivalents, limestones or
            dolostones).

   Class  5—Covered by glacial debris or Quaternary
            alluvial material  which  obscures  the
            bedrock.

These bedrock classifications were based on recent
state geologic maps generally  of 1:250,000-scale or
1:500,000-scale.  The  bedrock classifications were
developed  by Norton  et  al. (1982)  to evaluate
regional-scale sensitivity to  acidification. For this
comparison, a class was assigned to each lake based
on the dominant bedrock class in  which the lake
basin was  located. Presumably,  lake chemistry
represents the integrated  contribution of all  rock
types within a watershed, and  it is possible that rock
types (e.g., calcites or dolomites)  which  may
represent a  relatively small proportion of the total
watershed bedrock could have a dominant influence
on the lake chemistry. Thus, a formal test of the
utility of  these maps would require that the class
assigned to a lake be based on the synthesis of all
bedrock classes within the total watershed, not just
within the lake basin as presented here. Neverthe-
less, the  bedrock classifications for several  subre-
gions are sufficiently homogenous to be acceptable
for this application.

Population estimates for the proportion of low ANC
(<50//eq L~1) lakes by bedrock class show reasonably
good agreement  between the potential sensitivity of
the lake  as indicated by ANC and bedrock class
(Figure 6-21). In most subregionsthe lakes in bedrock
Class 1  (low to no acid neutralizing capacity) have
the highest proportion of low ANC lakes, whereas
lakes in the less sensitive bedrock classes showed
low  proportions of low ANC  lakes. This pattern
showed inconsistencies in California and the Central
Rockies. This may be an artifact of the criterion used
here to define low ANC lakes. Nearly all of the Wind
River Range was in bedrock Class 1, yet a small
percentage of lakes in this area (6.2%) had ANC <
50 //eq L"1. However, over 89 percent of the lakes
in the Wind Rivers had ANC < 200 //eq L"1. Using
a different criterion to define sensitive bedrock types
(such as ANC  < 200 //eq L~1) would  result  in  a
consistent ANC-bedrock  pattern  in the  Central
Rockies. The  pattern for  Ca+2 (<  50 //eq L"1) was
similar to that  presented  for ANC, whereas  little
relationship to the bedrock classes was observed for
conductance and SiOa.

The  importance  of  evaluating the influence of
bedrock type within subregions is apparent from the
large differences in the proportion  of low ANC lakes
observed among subregions within the same bedrock
class. The  higher proportion of low ANC  lakes in
California  and the Pacific  Northwest, compared to
the Central and Southern Rockies on similar bedrock
classes, likely reflects additional  factors, such as
precipitation and evaporation (Section 6.6).

6.4  Physical Lake Characteristics and
Water Chemistry

To examine more closely the relationship between
physical and  chemical lake characteristics in the
West, the subregions were divided into geomorphic
units. These units represent physiographic areas of
common geologic origin, the boundaries of which
generally coincide with the mountain ranges in the
West (Figure  6-22). Evidence for common  geologic
history (Hunt  1974; Chronic 1986) and topographic
features were used to delineate geomorphic units,
which generally follow  those in Snead (1980).
Although  population estimates were computed for
some geomorphic units in which  10 or more lakes
were sampled, extensive descriptions are given only
for those in which the sample size was at least 25.
Qualitative descriptions  are  presented for  those
subpopulations based on a sample size of 10 to 25.
Population estimates for selected geomorphic units
are presented in Section 6.4; the complete results
for geomorphic units are discussed in Section 7.

6.4.1  Lake Area

The major distinction between the target population
of lakes defined for the WLS-I and that defined for
the ELS-I  is that the population of western lakes
included those < 4 ha (but > 1 ha) in size (Section
2.2.1),  whereas the ELS-I sample  lakes  were
                                                                          105

-------
Figure 6-2 J.   Percentage of low ANC (< 50
             fieq i~V lakes by bedrock class
             {Norton et al. 1982) in Cali-
             fornia  (4A). the  Pacific
             Northwest (4B), the Northern
             Rockies (4C), the  Central
             Rockies (4D), and the South-
             ern Rockies (4E). Western
             Lake Survey - Phase I. Bed-
             rock sensitivity is highest in
             Class 1 and lowest in Class 3.
50


7 40
of
0
10
VI
o
«; 30
^
1
in
•5 20
1
10
0


—





	



-
-
— I
























-
__
^
N
s
1
^
1
I
W////Z7////7///77////////?.






























\

, 	 :







Y/&/&////Z&//A
Bedrock Class
D (Highest
i Sensitivity)
D-

H! 3
g^i


	




J
I









-

"/////////////////////////A
-
                                                   \        \       \        \        \       \
                                               California  Pacific  Northern  Central  Southern   West
                                                       Northwest Rockies  Rockies  Rockies
restricted to those > 4 ha (Linthurst et a\. 1986).
Small lakes (those  >  "\  and < 4 ha) comprise a
significant percentage of the western lake resource,
representing 43.8 percent of the target population.
The  relationship between  lake  area and  water
chemistry in western lakes was evaluated to examine
the physical and chemical characteristics of low ANC
lakes in  the West and to determine if population
estimates in the WLS-I needed to be restricted to
lakes > 4 ha for later comparison to eastern lakes.
In this analysis, large lakes are defined as > 30 ha
and small lakes as < 30 ha.

Lake area and ANC, SO*'2,  and Ca+2 showed little
relationship, as indicated by the correlation coeffi-
cients (Table 6-10). The lack  of a strong relationship
between ANC and lake area is apparent for the Pacific
Northwest (Figure 6-23). However, because there are
relatively few  large lakes (> 30 ha) in the Pacific
Northwest, no pattern can confidently be  deter-
mined. For lakes < 30 ha,  which  represent 92.4
percent of the target population for the West, no
relationship between  lake chemistry and lake area
was observed. However, western lakes > 30 ha had
higher concentrations of ANC  compared to those
lakes < 30 ha (Figure 6-24).

The impact of including lakes < 4 ha in the population
estimates for the WLS-I was evaluated by comparing
         quintiles (Qi and Q4) and median values for ANC,
         Ca+2, and S04~2 for lakes > 4 ha and lakes < 4 ha
         (Figure 6-25). Only in the Northern Rockies did lakes
         < 4 ha have substantially  lower concentrations of
         ANC, Ca+2 and S04~2 than lakes > 4 ha. Cumulative
         frequency distributions for ANC in four subregions
         showed little difference between lakes > 4 ha and
         <4 ha; only in the Northern Rockies was there strong
         indication that the  chemistry of these two groups
         of  lakes differ  (Figure 6-26).  Within  the Northern
         Rockies, a high percentage of small lakes are located
         in the geomorphic units overlying the Idaho Batholith
         (Bitterroot, Sawtooth, Clearwater, and Salmon River
         Mountains). This area also contains most of the low
         ANC lakes within the Northern  Rockies. Thus, the
         apparent positive relationship between ANC and lake
         area within the Northern  Rockies is  found  to be
         spurious in this case when the analysis  is focused
         on the geomorphic unit. Because the small lakes
         in the Northern Rockies represent only 10 percent
         of  the estimated number of lakes in  the western
         target population,  their impact on  the regional
         population estimates was minimal.

         In summary, lakes with areas > 1 and < 4 ha and
         lakes > 4 ha in four subregions of the West showed
         no appreciable  differences in ANC, Ca+2, and S04~2.
         A substantial chemical difference between large and
         small  lakes was observed only in the  Northern
                     706

-------
Figure 6-22.
               Locations and names of geomorphic units used in data analysis for specific subpopulations of lakes. Western
               Lake Survey -  Phase I, These units represent physiographic areas  of common geologic origin and generally
               coincide with major mountain ranges.
        Olympic
        Uplift
                   North
                   Washington
                   Cascades
                                              4C
                                  Selkirk     Cabinet
                                  Mountains  Mountains
  Puget
  Lowlands
Klamath
Mountains
     Middle
     Washington
     Cascades
    Wenatchee
    Mountains
South  I
                                                    Bitterroot
                                                    Range   \
                                                       Anaconda-Pintlar
                                                       Mountains      Beartooth
                        Washington   ,,lAwater
                        Cascades __,w_A_^tns-
                                  Salmon River
                                  Mountains
                                       Sawtooth
                                       Mountains
                                                                                        Medicine Bow
                                                                                        Range
                                                                                                4E
                                                                                             Front
                                                                                             Range
                                                                                             Sawatch Uplift
                                                                                              Sangre De Cristo
                                                                                              Uplift
  Rockies, where the difference was apparently caused
  by a spatial separation of large and small lakes into
  areas with  contrasting  bedrock  characteristics.
  Finally, the  number of lakes with  area < 4 ha in
  the  Northern  Rockies  is insufficient to  seriously
  influence the regional population estimates for the
  West. On  the  basis of this analysis, there appears
  to be little reason to eliminate the small lakes from
  the western population estimates when comparing
  them to eastern lakes, despite the difference in the
  sampling frames for the two surveys.
                                                     6.4.2  Lake Elevation

                                                     An obvious distinction between lakes in the East and
                                                     lakes in the West is that the elevation  of western
                                                     lakes was generally high and exhibited  substantial
                                                     subregional variation. Based on previously collected
                                                     data, it was assumed that most low ANC lakes in
                                                     the West would be located at high elevation  sites
                                                     (Omernik and  Griffith 1986; Turk 1983; Gibson et
                                                     al.  1983).  In  the  West, an  inverse relationship
                                                     between ANC and elevation was found in the Central
                                                                                107

-------
Table 6-10.    Spearman Correlation Coefficients (ra) for Lake
             Area versus ANC,  Calcium, and Sulfate by
             Subregion  and Stratum, Western  Lake
             Survey - Phase I
Correlation Coefficient
Subregion
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies
Stratum
4A1
4A2
4A3
4B1
4B2
4B3
4C1
4C2
4C3
4D1
4D2
4D3
4E1
4E2
4E3
n
54
53
42
59
53
47
53
50
40
43
47
39
46
52
41
ANC
-0.011
0.060
0.216
0.016
0.216
0.023
0.051
0.230
0.330
0.001
0.081
-0.080
-0.338
0.041
0.021
Ca+2
0.019
0.133
0.221
0.065
0.162
-0.129
0.000
0.256
0.342
0.108
0.060
-0.054
-0.292
0.056
0.009
so4-2
-0.019
0.376
0.199
0.206
0.132
-0.022
0.033
0.321
0.246
-0.307
-0.087
-0.041
-0.068
0.091
-0.075
3 Regression was performed using weighted values of variables,
 i.e., the values are population estimates (Section 5.1.2.1).
Rockies and in California, but the relationship was
very weak or  absent in the  Pacific  Northwest, the
Northern Rockies, and the Southern Rockies (Figures
6-27 to 6-31). The absence of a strong relationship
between ANC  and lake  elevation in the  Northern
Cascades also has  been reported  by Brakke and
Loranger (1986).

Rank correlation  coefficients within  geomorphic
units indicate  a  somewhat stronger degree  of
association between  ANC  and  elevation than is
observed within subregions (Table 6-11),  although
the relationship is still weak. In the Park Range and
the Uinta Mountains, the presence  of one outlier
in each unit  greatly influenced  the correlation
coefficient. In the Middle Washington Cascades, the
positive slope and the correlation coefficient suggest
that  lake  ANC  increases  with elevation.  This
relationship again is observed to be spurious because
of the presence of two  high ANC  lakes at  high
elevation. Only in the Wind River Range was the
decreasing relationship between ANC and  elevation
substantially demonstrated.

The relationship between  ANC and  lake  elevation
was explored  further by examining high  elevation
Figure 6-23.    Relationship between ANC
              {< 1000 peg L~V and lake area
              (<  100 ha) for lakes in the
              Pacific Northwest   (4B),
              Western Lake Survey - Phase I.
              The alkalinity map class from
              which each lake was selected
              is shown fD: < 100 fjeg L'\-
              A:  100-199 tieq L~\ X 200-
              400 [Jeq i~1; Omernik and
              Griffith 1986).







^
'fj
Cr
O



900-

800 _

700-
600-

500-

400.
300-

200-
1OO -
0
Alkalinity Map
Class foeq L~1J
n 
-------
 Figure 6-24.    Estimated number of lakes with ANC < 50
              lieq i'1 or ANC > 50 peg i"' by lake area
              class. Western Lake Survey - Phase I.
       25
       15
   vi
    6000
    5000
  * 4000
  -J
  ^ 3000


  "2 2000 -

  I
  <2 rooo
                       Acid Neutralizing
                       Capacity (peq L'^j
                         n >5°
                             <50
          0    5    10   15   20   25   30
                   Lake Area Class (ha)


lakes, defined as those at elevations exceeding 3000
m. An elevation of 3000 m generally corresponds
to timberline,  although  timberline  varies from
approximately 1800 m in the Olympic Mountains of
Washington to over 3600 m in the Sangre de Cristo
Mountains of Colorado because of temperature,
wind, and other factors (Arno and Hammerly 1984).
No  lakes  > 3000  m were sampled in the  Pacific
Northwest or in the Northern  Rockies. In California
and the Central Rockies, high elevation lakes have
slightly lower concentrations of ANC and Ca*2 than
lakes at low elevations (Table 6-12). In the Southern
Rockies,  however,  high elevation  lakes  have  far
lower concentrations than those  at low elevations.
The  minimum elevation of lakes  in the Southern
Rockies is 2100 m; thus, it is surprising that an
elevational gradient of 900 m  would result in such
large differences in ANC, Ca+2, and S04'2.

A previous example (Section 6.4) showed that the
apparent relationship between lake area and water
chemistry in the Northern Rockies was attributed,
in part, to spatial separation of large and small lakes
into geomorphic units with different bedrock types.
 To illustrate  confounding  orographic effects  in
 evaluating  the relationship  between ANC  and
 elevation, the lakes in the Sierra Nevada and Oregon
 Cascades were grouped as those located east and
 west of the topographic divides for the respective
 ranges. Both mountain ranges have similar bedrock
 types  across  the  width of the units.  Population
 estimates for ANC, Ca+2, and S04"2 concentrations
 for lakes in the eastern Sierra  Nevada were much
 higher than for lakes on the western slope (Table
 6-13). In contrast, the chemistry of the lakes across
 the Oregon Cascades is homogeneous, as illustrated
 for ANC  (Figure  6-32). These observations are
 consistent with the pronounced "rain shadow" effect
 on the eastern slope of the Sierra Nevada indicated
 by a 3000-m crest throughout much of the mountain
 range. The crest of the  Oregon  Cascades  is
 approximately 1500  m less than that of the Sierra
 Nevada, with only isolated volcanic peaks approach-
 ing the elevation of the  Sierra Nevada  crest.
 Consequently, the orographic effects on ANC in lakes
 in the Sierra Nevada are striking, whereas little effect
 is observed in the Oregon Cascades.

 6.4.3   Hydrology

 6.4.3.1  Hydrologic Lake Type—A  qualitative
 assessment of the influence of hydrology  on the
 chemistry of lakes was performed by categorizing
 lakes  on the basis of the presence or absence of
 permanent inlets and outlets. The objective of this
 classification was to distinguish lakes that receive
 substantial   hydrologic   contributions   from
 watersheds from those that do not. Greater contact
 time with neutralizing processes in watersheds can
 increase the ANC of inflowing water. Thus, in lakes
 for which the  hydrologic contribution is primarily
 through the watershed, rather than  as a direct
 surface input,  lake ANC concentrations would  be
 expected to be higher. For the same reason, higher
 ANC concentrations would also be expected in lakes
for which the watershed area to lake area ratios are
 large. The presence of permanent inlets and outlets
was  determined by  examining large-scale USGS
topographic  maps. Lakes  were classified  (Wetzel
 1983) as seepage  (no inlets, no outlets), drainage
(outlet), closed (inlets, no outlets), and reservoir
(outlet control structure present).
For the western  United  States, the chemistry of
drainage lakes and seepage lakes was similar (Table
6-14).  In  California and the Pacific  Northwest,
drainage lakes had higher estimated concentrations
of ANC and Ca+2 than did seepage lakes. In the three
Rocky Mountain  subregions, the inverse was true.
Sulfate  and DOC concentrations in both drainage
and seepage lakes were low, but drainage lakes had
slightly higher S04~2 and lower DOC concentrations
                                                                          109

-------
Figure 6-25.   Concentrations f/jeq L'^) of ANC. sulfate and calcium in lakes <4 ha (	) and > 4 ha ( •
                                             ) in California
    7200 -
    7000 -
     500 -
     600 -
     400 -
     200 -
             (4A). the Pacific Northwest (48). the Northern Rockies (4C). the Central Rockies (4DJ, the Southern Rockies
             (4£), ^and the West (4) in the Western Lake Survey - Phase I. First quintiles (Qi. 20th percentile) medians
             (M, 50th percentile). and fourth quintiles (Q4. 80th percentile) are given for each variable (.'

-
-
.
-
.



•
_

c
9
_ 1
1
99 ;•
"xi 2
w ,
c









9

] .



1
i
9
0









«t"
9
1
1 LI
', | i
.
-
.



-
fl






, fo JJ ~° i,
— ii ii <
i i i
1



	


»

720

700

80
&
CD
^ 60
u
1
40



20



n
~ <4ha '
' Q, M Q4 (p
>4/J8
o | a
' Q, M Q4
-
-

-

9 o
1C

-

"~ n
p

?J- T,
YY A
p


9
1
1
1
!
*
iD
i I
•M
1 I
.1 .
Q*1^ O
p _
-
^
-
—
-

-

~
9n-

-

• > —
•f"
6
» A* i°



soo
700
600
f" 500
o-
f 400
.§
(0
 i
1 — l-l
]
ft







9
T
T
TjT J^
i AO ^^
1 	 M 	 1-

p




9
c




. T
^ 1
6<
-
-
-
«.
-
M

P
—
-
—



1 "
          4A  4B 4C 4D 4E  4
4A 4B 4C 4D 4E  4
4A  4B  4C 4D 4£  4
than  did  seepage lakes.  Lack of major chemical
distinctions between drainage and seepage lakes in
the West  contrasts with the large chemical differ-
ences between these lake types observed in the East
(Linthurst et al.  1986). This  may result from  the
highly seasonal nature of surface input into western
lakes; consequently,  intermittent  runoff, although
high in total volume from the deep snowpack, would
not be identified as a permanent inlet on a map.
Also, meltwater passing through highly porous talus
slopes would not be identified as an inlet. Thus, using
the presence of stream inlets from  mapped  infor-
mation may have limited utility in the West. Except
in the Pacific Northwest, reservoirs had relatively
high  concentrations  of  DOC and  SOT2. These
elevated concentrations in reservoirs are consistent
with large contributions from the watersheds.
 The Pacific Coast  subregions receive  far greater
 amounts of precipitation and experience lower rates
 of  evaporation  than  do the Rocky Mountain
 subregions (U.S.  Department of Commerce 1974).
 For lakes with equal watershed areas  and similar
 lake morphometries, higher  rates  of  precipitation
 result in greater dilution. Thus, the combined effects
         of greater precipitation  and lower rates of evapo-
         transpiration  might account for the lower concen-
         trations of chemical constituents observed  in the
         Pacific Coast subregions. Differences in precipitation
         and evaporation also occur on the slopes of the Sierra
         Nevada, which are reflected in higher values of ANC
         on the eastern slopes (Section 6.5).


         6.4.3.2  Hydraulic  Residence  Time—Use  of
         hydrologic lake types  is a qualitative approach for
         evaluating the relationship between lake chemistry
         and hydrology. A more quantitative approach is to
         compute  hydraulic  residence time.  Hydraulic
         residence  time is  defined as the time required to
         exchange the total volume of water in a lake:


         Hydraulic Residence Time =
            Lake Volume/Total Volume of Inflow per Year

         A large, deep lake with a relatively small watershed
         would be expected to have  a  long  (e.g., > 3  yr)
         hydraulic  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.
                      110

-------
Figure 6-26.     Cumulative frequency distributions [F(x>] for ANC (< 1000 peg L V *nd calcium (< 1000 fjeq L V- and inverse
                cumulative frequency distributions [1-F(x)] for sulfate (< 100 fjeq L~^) in lakes < 4 ha (	) and > 4 ha (	}
                in the Northern Rockies (4C), Western Lake Survey - Phase I. Other subregions showed no substantial differences.
                 FM
                        1.0
                        0.8-


                 FM    0.6 -


                        0.4-


                        0.2-


                        0.0
                        1.0
                        0.8 -
                        0.4 -
                        0.2 _
                        0.0
                                         200
 I
400
 I
600
 1
800
                                                                                                    1000
                                                           ANC (fjeq
                               <4ha

                               ^4 ha
                                         200
 \
400
                                                                        I
                                                                       500
               500
                             WOO
                                                          Calcium (fjeq L'^)
                        1.0 .
                        0.8 _


                1-FM    0.6 _


                        0.4 _


                        0.2 _
                        0.0
                                          20
  I              i
 40            60
   Sulfate (fjeq L'')
                              <4ha

                              > 4 ha
                                                                                       I
                                                                                      80
                             100
                                                                                         111

-------
Hydraulic  residence time was approximated using
the following expression:

             RT = LA (0.464 ZS)/[RO (WA - LA) +
                  LA (PRECIP)]

where:

             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),  interpo-
                  lated from  National  Weather
                  Service data.

Lake volume was computed by multiplying lake area
by mean depth, the latter of which was approximated
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 and may deviate considerably for individual
lakes.  Runoff was  estimated for  each lake's
catchment using linear interpolation from mapped
contours  of mean  annual  runoff  (Busby 1966).
          Precipitation at each  lake was interpolated from
          precipitation maps  (Environmental Data  Service
          1983). Some of the assumptions and potential errors
          in estimating hydraulic  residence  time with this
          method are discussed  in Linthurst et al. (1986). An
          additional  source of error in  computing hydraulic
          residence times, especially for western lakes, is the
          assumption that  surface runoff completely mixes
          with water present in  the lake. This assumption is
          violated where snowmelt runoff flows through the
          lake to the outlet without appreciably mixing with
          the denser lake waters.  This might be expected to
          occur more commonly in deep lakes with relatively
          small surface areas. Lakes responding in this manner
          will  have  a longer hydraulic  residence time than
          indicated  by the estimate calculated here. Another
          problem with use of mapped values for estimating
          hydraulic  residence time is that orographic effects
          can  cause high spatial  variability  in precipitation
          amounts.

          The  population estimates  for hydraulic  residence
          time  among subregions (Table  6-15)  are  similar,
          except for  the Pacific Northwest. In this subregion,
          which experiences  high  total  precipitation and
          relatively  low  evapotranspiration  rates, hydraulic
          residence time is approximately 50 percent of those
          in other   subregions. Although the  estimated
          hydraulic  residence times for California and the
Figure 6-27.    Relationship between ANC
              (< 1000 iteq L"1! and eleva-
              tion (m) for lakes in California
              (4A). Western Lake Survey -
              Phase I. The  alkalinity map
              class from which each lake
              was selected  is shown ( D:
              < 100 peg L'\- A; 100-199
              Vieq L"; X200-400 fjeq r1;
              Omernik and Griffith 1986).
300.


800-


700


600


500 _
                                          300 _


                                          200-


                                          700 _


                                           0
Alkalinity Map
Class ffjeq L'^j
n  < 100

A  100 - 199
X  200-400
                                                    i      i      i      i       \      \      i
                                                   500   1000   1500   2000  2500  3000  3500  4000

                                                                  Elevation (m)
                     112

-------
Figure  6-28.    Relationship between ANC
               (< 1000 fjeq i~V *nd eleva-
               tion (m) for lakes in the Pacific
               Northwest  (4B).  Western
               Lake Survey - Phase I.  The
               alkalinity  map class from
               which each lake was selected
               is shown ( D: < 100 fjeq L'1;
               A. 100-199 fjeqL~\ X: 200-
               400 peq i~1;  Omernik  and
               Griffith 1986).




^
ANC (fjeq L'



1000
900 -
800-
700-
600 _

500-
400 _
300 _
200 _
700 _
0

Alkalinity Map
Class (fjeq L''1)
D < 700
A 700 - 799
X 200-400
„
a n
o
x y
a D
X
ax x
A
4x i a
& °x& C? fi * ^! a
* x D *x
°°° ^rf,^" Dc
i i 1 t i 1 1
                                                     0     500    7000   7500  2000   2500  3000   3500  4000
                                                                            Elevation (m)
      1000
    900 _


    SOO _


    700 _


    600


T-J  500
 o-

 O  400


    300 _


    200 _


    700


      0
Alkalinity Map
Class (fjeq L'^)
D  <700
A  700-799
X  200 - 400
                                            P*°£D  D

                                            ]Bn  «^*
                                                                       Figure 6-29.    Relationship  between ANC
                                                                                      (< 1000 iieq O and eleva-
                                                                                      tion (m) for the Northern
                                                                                      Rockies <4C). Western Lake
                                                                                      Survey - Phase I. The alkalin-
                                                                                      ity map class from which each
                                                                                      lake was selected is shown
                                                                                      (O: < 700 fieq L~\-  A: 700-
                                                                                      199 fjeq L'1; X. 200-400
                                                                                      fjeq i"1; Omernik and Griffith
                                                                                      1986).
            0    500   1000  1500  2000  2500   3000   3500   4000
                                 Elevation (m)
                                                                                      113

-------
Figure 6-30.    Relationship between ANC
               (< 1000 /jeq i"V and eleva-
               tion (m) for the Central Rock-
               ies <4D).  Western  Lake
               Survey - Phase I. The alkalin-
               ity map class from which each
               lake was selected is shown
               (O: < 100 neq L~'; A: 100-
               199 fjeg L'\- X:  200-400
               fieq i"1; Omernik and Griffith
               1986).
1000


 900.


 800-


 700.


 600-


 500.


 400_
                                               200-,


                                               100


                                                 0
Alkalinity Map
Class (ueqL^)
D  <100
A  100 - 199
X 200 - 400
                                                          i       i       i       i       i        i       r
                                                         500   7000  7500  2000   2500   3000   3500   4000
                                                                         Elevation (m)




'-J
0-
J




900-
800-
700-
600-
50O-
400-
300.

200
100
0
Alkalinity Map 5
Class tueq O
D <700
A 100 - 199
X 200-400
\ '
X


x a
fl a
^ A
D x° fl
&
A X 6 y
X6
jp D a ^a x
" ° D^ »
i i i i i i i
                                                                        Figure 6-31.
                                         Relationship between  ANC
                                         (<  1000 neq L'^> and eleva-
                                         tion (m)  for the Southern
                                         Rockies (4E).  Western Lake
                                         Survey - Phase I. The alkalin-
                                         ity map class from which each
                                         lake was  selected is shown
                                         ( Q < 100 peq L~\- A: 100-
                                         199 veq  L'\- X: 200-400
                                         peg L~\- Omernik and Griffith
                                         1986).
                   500    1000   1500   2000   2500   3000   3500   4000

                                     Elevation (m)
                        114

-------
Table 6-11.    Weighted Correlation Coefficients for ANC versus Lake Elevation by Geomorphic Unit, Western Lake Survey
              Phase I
Subregion
California
Pacific Northwest





Northern Rockies




Central Rockies

Southern Rockies


Geomorphic Unit
4A Sierra Nevada
4B Oregon Cascades
South Washington Cascades
North Cascades
Wenatchee Mountains
Middle Washington Cascades
North Washington Cascades
4C Idaho Batholith
Salmon River Mountains
Sawtooth Mountains
Bitterroot Range
Clearwater Mountains
4D Wind River Range
Uinta Arch
4E Park Range
Front Range
San Juan Mountains
n***
114
42
20
68
32
24
12
88
22
20
37
g
47
27
21
49
21
Correlation
Coefficient
(r)
-0.232
-0.298
-0.053
-0.104
-0.242
0.469
-0.280
0.156
0.302
0.159
-0.395
0.230
-0.582
-0.448
-0.922
-0.429
-0.586
Table 6-12.    Population Statistics for ANC, Calcium, and Sulfate in High (> 3000 m) and Low (< 3000 m) Elevation Lakes
               in California (4A). the Central Rockies (4D), and the Southern Rockies  (4E), Western Lake Survey -
               Phase I
                     ANC (^eq L~1)
                                                                         Ca
                                                                           +2
Elevation Category
                                                  M
                                                                             M
                                                              cx»
                                                                                                      M
California

  > 3000 m
  < 3000 m
                (4A)
Central Rockies   (4D)

  > 3000 m
  < 3000 m

Southern Rockies (4E)
                         38
                        111
73
56
      1215
      1186
                               1363
                                936
35.9
29.6
59.7
88.5
                                                  54.5
                                                  65.6
 92.5
166.9
106.8
204.5
128.8
464.5
29.4
19.9
61.4
89.1
                                   43.0
                                   43.2
                                                                            83.3
                                                                           140.4
                                     85.1
                                    126.6
                                             112.0
                                             333.9
                                     4.9
                                     2.3
                                    20.1
                                    15.7
                                  7.3
                                  4.6
                                 23.9
                                 26.3
26.7
10.5
30.9
38.7
> 3000 m 118 1110
< 3000m 21 499
n*"*
4V
N
QI
M
Q4
= number of lakes from which samples
= estimated number of lakes.
= first quintile (20th percentile).
= median (50th percentile).
= fourth quintile (80th percentile).
76.5 174.0 567.0 70.0 141.2 410.7 14.0 27.3 47.4
449.0 1116.4 1673.8 386.3 810.6 1235.4 29.9 78.6 206.3
were obtained.
                                                                                       115

-------
Table 6-13.    Population Statistics for ANC, Calcium, and Sulfate in Lakes on the Western and Eastern Slopes of the Sierra Nevada
              and Oregon Cascades, Western Lake Survey - Phase I

nrionta+irtrt
Mountain Range of Slope n*** N Q,
Sierra Nevada
West 88 1697 28

Oregon

n*** =
N
Q-i =
M =
Q4 =
Figure







East 26 422 86
Cascades
West 22 231 38
East 14 148 30
number of lakes from which samples were
estimated number of lakes.
first quintile (20th percentile).
median (50th percentile).
fourth quintile (80th percentile).
6-32. Relationship between ANC
(< 500 iieq i~1J and elevation
(m) for lakes on the western
and eastern slopes of the
Sierra Nevada in California
(4A) and Oregon Cascades in
the Pacific Northwest (4B),
Western Lake Survey -
Phase 1.







ANC (/jeq L""1) Ca+2 (jieq L~1) SO4~2 (j*eq L~1)
M Q4 a, M 0.4 Q, M Q4
9 52.2 94.4 20.5 36.0 64.5 4.0 6.4 10
9 138.4 289.6 66.2 116.9 378.0 5.3 26.0 91
9 91.5 166.0 17.2 48.9 85.5 0.9 1.8 4
4 66.3 209.4 8.0 30.4 88.5 0.4 1.8 3
obtained.
^nrt Sierra Nevada
West Slope fast Slope
400 . .
,- 300 ' .
-j
I " "
•2: 200 •
* ' . . •
<* \-
100 •,•.•••/•
' ' \ .\'-':4r: • ••' ' •
o •«••••
7000 2000 3000 4000 7000 2000 3000 4000
Elevation (m)
Oregon Cascades
hOO
West Slope East Slope
^400
J1300
^~ •
^ • .
^200 .
. • *
n
500 7500 2500 500 7500 2500
Elevation (m)
                          116

-------
Table 6-14. Population Medians for Selected Parameters by Hydrologic Lake Type, Western Lake Survey Phase - I
Subregion
California (4A)



Pacific Northwest (4B)



Northern Rockies (4C)



Central Rockies (4D)



Southern Rockies (4E)



West (4)



Lake Type
Drainage
Seepage
Reservoir
Closed
Drainage
Seepage
Reservoir
Closed
Drainage
Seepage
Reservoir
Closed
Drainage
Seepage
Reservoir
Closed
Drainage
Seepage
Reservoir
Closed
Drainage
Seepage
Reservoir
Closed
n**
107
27
14
1
115
35
7
2
118
18
4
3
88
31
7
3
104
21
13
1
532
132
45
10
N
2013
282
97
8
1224
380
80
22
1916
292
92
80
1592
551
108
48
945
322
337
4
7691
1826
714
163
WA:LAa
23.1
15.2
24.1
_ b
20.9
10.8
—
—
20.4
26.0
—
—
32.6
20.0
—
—
24.1
13.2
91.2
—
24.0
16.6
75.7
24.2
Site Depth
(m)
9.1
6.5
12.6
—
8.7
4.0
—
—
9.4
6.1
—
—
8.6
7.8
—
—
5.3
4.3
5.0
—
8.7
5.8
7.7
10.4
Lake Area
(ha)
4.7
2.7
37.3
-
4.6
3.2
—
-
4.6
3.1
—
-
5.5
3.4
—
-
3.1
4.1
5.1
-
4.6
3.4
15.6
9.0
ANC
61.8
54.1
142.9
-
148.5
79.6
—
-
178.5
271.0
—
-
98.4
128.9
—
-
180.9
201.6
1051.6
-
108.1
113.2
554.1
196.0
DOC
(mg L"1)
1.82
1.74
1.28
-
1.13
2.09
—
-
1.15
1.12
—
-
1.09
1.60
—
-
1.06
2.19
3.55
-
1.06
1.84
2.53
1.85
Ca+2l
43.4
36.4
86.6
-
103.9
41.0
—
-
146.2
209.1
—
-
89.7
111.4
—
-
178.2
141.2
722.0
-
88.3
70.4
422.0
97.4
so4-2
6.7
4.8
6.3

17.1
5.6
—
-
13.5
12.3
—
-
24.3
21.8
—
-
28.7
24.5
77.8
-
17.8
16.6
49.6
17.2
 n***    = number of lakes from which samples were obtained.
 N      = estimated number of lakes.
"WA:LA  = Ratio of watershed area to lake area
b(—) Insufficient sample size, n*** < 10.
Table 6-15.    Estimated Hydraulic Residence Time (Years)
             for Drainage Lakes and Reservoirs by Subre-
             gion, Western Lake Survey - Phase I
                                      Hydraulic
                                  Residence Time (yr)
Subregion
California
Pacific Northwest
Northern Rockies
Central Rockies
Southern Rockies

4A
4B
4C
4D
4E
n***
121
122
122
95
117
N
2110
1304
2007
1700
1282
Qi
0.10
0.05
0.13
0.06
0.06
M
0.29
0.16
0.42
0.29
0.34
Qi
0.97
0.57
1.33
1 16
1.02
West
577   8403   0.08   0.29   1 12
n***   = number of lakes from which samples were obtained
N     = estimated number of lakes.
Qi    = first quintile (20th percentile).
M     = median (50th percentile).
Q4    = fourth quintile (80th percentile)
Southern Rockies lakes are similar, the greater depth
of the California lakes compensates for the higher
precipitation  received compared  to  lakes  in  the
Southern Rockies.

The relatively short residence times for lakes in the
West compared  to those in the East (Linthurst et
al. 1986) have  important  implications  relative to
potential impacts from atmospheric  deposition.
Lakes in the West likely receive most of their annual
inflow during snowmelt. Replacement of the lake
volume  in a short period of time would minimize
the opportunity for the runoff to be neutralized by
contact  with the  catchment  or  through  in-lake
processes. Thus, a high volume of snowmelt inflow
over bedrock that is  highly resistant to weathering
would explain why such a high proportion of western
lakes are so dilute.
                                                                               7/7

-------
                                          Section 7
                                Subregional Characteristics
7.1   California (4A)

California  (Subregion 4A) contained the  highest
percentage (36.7%) and the  highest number (880)
of western lakes with ANC < 50 //eq L~1 (Table 5-9).
Lakes  in  California  generally  had the  lowest
concentrations of Ca+2, SO.T2 and dissolved organic
carbon of any subregion (Table  5-16).  Secondary
variables for which California lakes had the lowest
median (M) values were total phosphorus, true color,
turbidity, sodium, potassium, magnesium, chloride,
conductance (Tables 5-17 to  5-19, Table 5-21), and
fluoride (Section 6.2.1). These very dilute waters in
California were the clearest lakes in the West, with
a median Secchi disk transparency of 7.4 m (Table
5-18).

In California,  58.2  percent of the  lakes had
conductance values < 1OfjS cm"1, representing more
lakes  in this class (Nc = 1398)  than  in all other
subregions combined  (Table  5-24). The bedrock is
similar to that in some of the areas in the  Rockies
containing batholiths, but it is distinct from that in
the Pacific Northwest (4B).  In addition, California
receives  slightly less total  precipitation than  the
Pacific Northwest (U.S. Department of  Commerce
1974), primarily as snow. The excess precipitation
(precipitation - evaporation) is  much  greater  in
California than  in the Rockies, thus contributing to
less concentrated solutes in lakes in California.

The lakes in California were the most dilute  and had
the lowest ANC concentrations, and also were  the
largest and among the deepest of any sampled in
the West. This  observation  contrasts with  that for
the East where the  lakes  with low  ANC were
generally much smaller than those with high ANC
(Linthurstetal.  1986).

Over one-half of the lakes in California were found
at elevations > 3000 m (Table 6-12). This proportion
was less than that for the Central Rockies (4D) and
the Southern Rockies (4E),  but was much  greater
than in the  Pacific  Northwest and the Northern
Rockies (4C) where no lakes > 3000 m were sampled.
The estimated  percentage of lakes  in  wilderness
areas in California (44.6%) was comparable to that
for  the  other subregions, except for  the  Central
Rockies where  the percentage of wilderness lakes
was 65.4 percent. Thus, the high percentage of low
ANC lakes in California cannot  be attributed to an
association of low ANC lakes with wilderness areas.

The geomorphic units (Section 6.3.2) for California
(Figure 7-1) were delineated to exclude lakes located
in the foothills. These lower  elevation lakes are
associated with  different vegetation,  soils and
landuse factors that are atypical of lakes located in
the mountainous areas. Over 91 percent of the target
population of lakes in California was represented by
three geomorphic units,  the  Sierra Nevada, the
Klamath Mountains and the California Cascades. The
minimum elevation of lakes selected in the geomor-
phic unit identified as the Sierra Nevada was 1219m
(4000feet). The minimum elevation of lakes selected
in the  California  Cascades and  in  the Klamath
Mountains was 914 m (3000 feet).

7.1.1   Sierra Nevada
The Sierra Nevada geomorphic unit extends approx-
imately 700 km and is about 100 km wide. The range
is dominated  by a  large  batholith  composed of
granites and granodiorite, although the northern and
eastern portions  contain  some metavolcanic and
metasedimentary rocks (Chronic 1986). The individ-
ual plutons comprising the Sierra  Nevada batholith
contribute to slight differences in  rock type; locally
dramatic differences  in  rock composition can be
attributed to the presence of dikes. Extensive faulting
resulted in the uplifting  of the eastern portion of
the Sierra Nevada. Consequently, the range is highly
asymmetrical  with the  western slope relatively
gentle,  in contrast to the extremely  steep eastern
escarpment. Most of the lakes in the Sierra Nevada
are located on  the broad western slope and annually
receive up to 400 cm of  precipitation. The amount
of  precipitation on  the  western slope is highly
dependent on  elevation. Average annual precipita-
tion increases by approximately one cm for every
30-m increase in elevation (Hunt 1974). On the
eastern slope, precipitation decreases dramatically
with  decreasing  elevation. The eastern  slope
generally receives less  than  50  percent of the
amount  of precipitation  received on the  western
slope  (Hunt  1974). The relationship  between
elevation and  ANC  for the eastern and  western
slopes of the Sierra Nevada is shown in Figure 6-29.
                                               118

-------
Figure  7-1.    Locations of sampled lakes within geomorphic units in California (4A), Western Lake Survey - Phase I. Lakes
              located outside the boundaries were not included in estimates for these subpopulations.
                                                                  OR
               Klamath
               Mountains
                                                                                                            \
                                                                                       119

-------
An  estimated 2119 lakes comprised the subpopu-
lation of lakes in the Sierra Nevada geomorphic unit.
Concentrations of all measured chemical constitu-
ents were  extremely low  in  Sierra Nevada  lakes
(Tables 7-1 and 7-2). The Sierra Nevada contained
an  estimated 834 lakes (39.3%) with ANC  < 50
//eq L"1. This large number, in part, reflects the large
surface area of the Sierra Nevada that was covered
during the WLS-I. However, the California, Oregon,
and Washington Cascades, a  range of comparable
surface area, contained fewer lakes with ANC < 50
//eq L"1 (351 or 24%, Tables 7-1 and 7-3). Over 88
percent of the lakes in the Sierra Nevada had ANC
values < 200 //eq L~\ The median ANC (59.8 //eq L"1)
was the lowest among lakes in all geomorphic units
examined in the WLS-I. A study of 26 high elevation
lakes (5000-9000 feet) on the western slope of the
Sierra Nevada in  1980-81  showed a much higher
median alkalinity concentration, exceeding 100//eq
L"1 (Tonnessen 1983).

Concentrations  of Ca+2 were similar to  those
measured for ANC. The median Ca*2 concentration
was 42.5 //eq L'1 and 61.0 percent of the lakes had
Ca+2<50//eqL"1.

Most lakes in the Sierra Nevada had extremely low
concentrations of DOC;  82.9  percent of the  lakes
had DOC < 2 mg L"1 (Table 7-1), and only 3.1 percent
had DOC > 6 mg l_"1 (Volume II). Similarly, very low
concentrations of S04~a were  found in lakes in the
Sierra Nevada. The median SO*"2 concentration was
7.1  //eq L~1, and only 8.6 percent of the lakes had
SO4 2 concentrations > 50 //eq L"1 (Table 7-1).
The total ionic concentration of Sierra Nevada lakes,
as reflected in conductance values, shows that these
lakes as a group are the most dilute sampled to date
in the United States. The total ionic concentration
of lakes in the Sierra Nevada  is comparable to that
measured in very low ionic strength lakes in Norway
and  Sweden (Henriksen and Brakke 1986). The
estimated median conductance of 8.2  //S cm"1
compares closely with that reported by Fox et al.
(1982) for 124 lakes in Sequoia, Kings Canyon, and
Yosemite National Parks in October,  1980. In sharp
contrast, the minimum conductance measured in the
eastern  United States was 7.8  //S  cm~1 and the
median  value for regions in the  East ranged from
38.5 //S cm"1 in the  Southern Blue Ridge to 90.8
//S cm"1 in Florida (Linthurst et al. 1986).

The median CI" concentration  (2.0 //eq L"1) in Sierra
Nevada  lakes was considerably less than  that
measured  in deposition at the  National Acid
Deposition  Program  (NADP)  monitoring sites  in
Yosemite and Sequoia National Parks (NADP 1984).
The annual CI" concentration, weighted for precip-
itation, measured at Yosemite was 4.5 //eq L"1 and
at Sequoia, 8.2 //eq L"1. Assuming  constant CI" in
precipitation between years,  the  fact that the lake
CI" concentrations do not reflect the composition of
the precipitation at these sites indicates that the
precipitation  measured  at  deposition  monitoring
sites is  not  representative of what the lakes are
actually receiving, or  that Cl~ is not conservative in
these watersheds. The elevation of the Yosemite site
is 1408 m and of the Sequoia site, 1856 m; both
elevations are considerably lower than those of the
Table 7-1.   Population Estimates and Statistics for Primary Variables by Geomorphic Unit in California (4A), Western Lake
           Survey - I
ANC < 50 fieq L"1
Geomorphic Unit n" N pc Nc Ncu
Sierra Nevada 114 2119 0.393 834 1033
California Cascades 16 102 0.402 41 64
Klamath Mountains 14 153 0.035 5 13
S04-*>50,eqL-'
Geomorphic Unit n*** N pc Nc Ncu
Sierra Nevada 114 2119 0.086 182 294
California Cascades 16 102 0.000 0 -
Klamath Mountains 14 153 0.000 0 -
n*'* = number of lakes from which samples were obtained.
N = estimated number of lakes.
ANC
QI
33.6
27.0
57.4
so4-
QI
4.1
0.7
2.5


(Meq
M
59.8
64.5
76.3

M
7.1
1.9
6.1


L"1)
Q4
126.2
176.2
286.1
1L-1)
Q4
13.7
3.2
10.1


Ca+2 < 50 /jeq L"1
Pc
0.610
0.592
0.035
DOC
Pc
0.829
0.485
0.744


NC Ncu
1292 1499
60 88
5 13
< 2 mg LT1
NC cu
1757 1933
49 76
114 192


Ca + 2
QI
22.4
9.0
46.7
DOC
QI
0.44
1.39
0.83


(Meq
M
42.5
22.8
59.5
(mg
M
0.83
2.04
1.30


L-1)
Q4
91.0
77.1
188.2
L-1)
Q4
1.80
2.57
2.08


pc = estimated proportion of lakes meeting the given criteria.
Nc = estimated number of lakes meeting the given criteria.
Ncu = 95 percent upper confidence limit for Nc.
Q! = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
( — ) = undefined.
















































                    720

-------
Table 7-2.   Population Statistics for Selected Variables for Lakes in Geomorphic Units in California (4A),
           Western Lake Survey - Phase I
Conductance
Mg+2
(/jeq L-1)
Ma +
K+ cr
SiO2
(mg L'1)
Fe
(H9 L-1)
Geomorphic Unit
                         M
Q,
                                     M
                                                  M
Q,
                                                              M
                                                                          M
Q,
                                                                                       M
                                                                                                  M
Sierra Nevada
California Cascades
Klamath Mountains
4.
3.
7.
9
8
9
8.2
8.2
11.0
Site Depth
(m)
Geomorphic Unit
Sierra Nevada
California Cascades
Klamath Mountains
Q
4.
2.
5.
1
9
6
2
M
8.9
5.4
9.7
3.4
15.8
5.4
6.3
30.2
11.3
Lake Area
(ha)
QI
2.2
2.0
2.8
M
4.6
4.0
4.1
10
6
9
.2
.4
.3
WA:
Q
12
9
10
1
.0
.2
.1
18.7
16.2
14.8
;LA3
M
22.8
15.4
23.7
2.6
2.6
3.3
3.6 1.4
3.5 3.1
4.5 3.3
2.0 1.24 1.99 4.8 13.8
3.6 0.29 1.06 14.3 19.5
5.0 1.26 1.89 17.0 34.0
Secchi Disk
Transparency
(m)
QI
4.2
3.4
5.4
M
7.5
5.8
6.8

Q-i = first quintile (20th percentile).
M = median (50th percentile).
aWA:LA = watershed area:lake area.
Table 7-3.   Population Estimates and Statistics for Primary Variables by Geomorphic Unit in the Pacific Northwest (4B),
            Western Lake Survey - I
ANC < 50 /jeq L"
Geomorphic Unit n*** N pc
Oregon Cascades 42 443 0.255
S. Washington Cascades 20 218 0.345
North Cascades 68 710 0.171
Wenatchee Mountains3 32 329 0.181
M.Washington Cascades3 24 248 0.169
N.Washington Cascades3 12 133 0.150
Puget Lowlands 22 252 0.089
SO.T2 >
Geomorphic Unit n*** N pc
Oregon Cascades 42 443 0.024
S. Washington Cascades 20 218 0.157
North Cascades 68 710 0.090
Wenatchee Mountains3 32 329 0.000
M.Washington Cascadesa 24 248 0.163
N.Washington Cascades3 12 133 0.177
Puget Lowlands 22 252 0.538
Nc
113
75
122
60
42
20
22
50 Aieq L
Nc
11
34
64
0
40
24
136
i
NCU
165
119
174
96
74
42
47
-1
NCU
28
65
105
—
72
50
195
ANC lf
-------
Figure  7-2.    Locations of sampled lakes within geomorphic units in the Pacific Northwest (4B), Western Lake Survey - Phase I.
              Lakes located outside the boundaries were not included in estimates for these subpopulations.
                                                       North
                                                       Washington
                                                       Cascades
                           Olympic
                           Uplift
                Puget
                Lowlands
              4B
                                                                South
                                                                Washington
                                                                Cascades
                                                                           Middle
                                                                           Washington
                                                                           Cascades
                                                                         Wenatchee
                                                                         Mountains
                                                                                            WA
                                                                                                           i
                                                                                                    i
                                                                                                    i
                                                                                            OR
                          122

-------
lower concentrations  of  ANC,  Ca+2, and Mg+2;
differences  in  CT  and S04~2  may,  in  part, be
attributable to differences  in analytical techniques
for these anions.

7.1.2  Other Geomorphic Units

As noted in the Survey design (Section 2), the desired
number of probability lakes  to be  sampled per
stratum was 50. Small departures from this desired
number do not substantially reduce confidence  in
the population estimates. However, substantial
reduction  in the achieved sample  size (e.g., n***
< 25) greatly  increases  the  uncertainty of the
population estimates. Subsequent to completion  of
the Survey, it was determined that stratifying the
data by  geomorphic unit was a useful approach  to
interpreting  patterns in lake  chemistry.  Some  of
these geomorphic units were of sufficient area (e.g.,
the Sierra Nevada) or contained a sufficient density
of lakes (e.g., the Wind River) such that the probability
sample  size  within these units  approached  or
exceeded the goal of 50 lakes per stratum. In many
geomorphic units discussed here, the sample size
was small and the confidence bounds for the
population estimates representing these units are
therefore greater.

The Klamath Mountains  (Figure 7-1) include the
Siskiyou Range, the Trinity Alps, and the Salmon
Mountains. This area contained relatively few lakes
(N = 153) and  too few were sampled (n*** = 14)
to have high  statistical confidence in population
estimates. The formation and composition of bedrock
in the Klamath Mountains is similar to that in the
Sierra  Nevada. Older metamorphic Paleozoic and
early  Mesozoic formations  were displaced by
intrusive granitic rocks in the Mesozoic era; these
formations were deformed by uplifting during the
Tertiary and Quaternary periods  (Hunt 1974). The
Klamath Mountains differ  from the Sierra Nevada
in that the former contains sizable areas of ultramaf ic
rocks. Of the  lakes in the Klamath Mountains, 96.5
percent had ANC values > 50 fjeq L~1. Nevertheless,
these lakes had a relatively low estimated median
ANC concentration (76.3 /ueq L"1). As was  observed
throughout the West, Ca+2 concentrations closely
paralleled those for  ANC. This suggests that either
the ultramafic  rocks were not  present  in these
watersheds or that they exerted little influence on
the ANC or Ca+2 concentrations.

Dissolved  organic carbon (DOC)  concentrations  in
lakes intheKlamath Mountains weresimilar to those
in the Sierra Nevada; all lakes sampled had DOC
< 6 mg L~1 (Volume II). Although the lakes in this
range are relatively close to the coast (the minimum
distance from the Pacific Ocean is about 75  km),
S04~2 and Cl~ concentrations were very low. All lakes
sampled had S04~2 concentrations < 50 fjeq L 1 and
the median concentration was 6.1  /ueq L~1.

Lakes in the Klamath Mountains are notably deep,
with a median site depth of 9.7 m, and have a high
watershed area to lake area ratio (M = 23.7). The
somewhat  higher concentrations of dissolved
material in these Klamath Mountain lakes may be
related  to their greater watershed size relative  to
those in the  Sierra Nevada and California Cascades.
A survey of  78 lakes in the Klamath Mountains by
Wilson  and Wood (1984) in summer and fall, 1983
showed a median conductance of 1 2 ^S cm'1 which
is close to the estimated value of 11.0  /vS cm~1
reported in this study. The median concentrations
of Ca*2 (37. 9 fjeq L"1) and Mg+2 ( 1 3.2 jueq L"1) reported
by Wilson and Wood (1984) also are close to the
concentrations reported here.

The California Cascades lie at the extreme  southern
end of the Cascade Range which extends northward
from  northern California  through Oregon  and
Washington. The extrusive bedrock of the California
Cascades contrasts to the intrusive coarse grained
granitic bedrock of the Sierra Nevada.  Despite this
major difference in bedrock type,  the lakes of the
California Cascades are very similar chemically to
those of the Sierra Nevada. The California Cascades
contain a higher percentage (40.2%) of lakes with
ANC < 50 fjeq L"1 than  does  the Sierra Nevada.
However, this percentage represents a small number
of lakes because relatively^ few lakes are present in
the California  Cascades (N  = 102, Table 7-1). Only
16 lakes were  sampled in this geomorphic unit, so
the statistical confidence in the population estimates
is not high. The Qi concentration of ANC was 27.0
peq L~1, almost  as low as that  for lakes  in the
Southern  Washington Cascades (21.3  fjeq L'\
Section 7.2.3). The Qi and median Ca+2 concentra-
tions for lakes  in the California Cascades  were the
lowest among  all geomorphic units (9.0 and 22.8
    L~1, respectively).
The  concentrations  of  DOC in the  California
Cascades (M = 2.04 mg L'1), although low relative
to those for eastern  lakes (Linthurst et al.  1986),
were approximately three times those observed in
lakes in the Sierra Nevada. However, the maximum
DOC  concentration  observed in  lakes  in  this
geomorphic  unit  was only  3.8 mg L"1.  Sulfate
concentrations  in  the California  Cascades  were
extremely  low, with a Qi value of 0.7 fjeq L"1 and
a median of 1 .9 fjeq L~1. Although numerous natural
sources of sulfur exist in this area (for example, hot
springs in Lassen  National Park), they apparently
have little impact on the chemistry of nearby lakes.
The S04~2 concentrations in these lakes and in the
nearby  Oregon Cascade  lakes are believed to  be
among  the  lowest in  the world  (Armstrong and
                                                                        123

-------
Schindler 1971). The lakes in the California Cascades
differ from those in the  Sierra Nevada in that the
California Cascades have much  higher  Mg+2 con-
centrations (indicative of the presence of mafic rocks
in the watershed), have lower SiO2 concentrations,
and are relatively shallow.

7.2  The Pacific Northwest

The size of the target population of lakes in the Pacific
Northwest (Subregion 4B) was estimated to be 1706
lakes. Among the subregions sampled in the West,
the Pacific Northwest contained the highest percen-
tage of lakes with pH < 6.0  (2.4%, Table 5-11) or
extractable aluminum > 50/ug L"1 (0.6%, Table 5-14),
and the lowest percentage of lakes with DOC > 6
mg L"1 (2.7%, Table 5-15). The  Pacific  Northwest
was second  only to California in the percentage of
lakes with ANC  < 50 pteq L"1 (19.5%, Table 5-9) or
Ca+2 < 50 /ueq L"1 (28.4%, Table 5-13). Concentra-
tions of ANC and Ca+2 tend to decrease from north
to south in this subregion.

The median  Secchi disk transparency (4.9 m, Table
5-18) was considerably less than that in California
(M = 7.4 m), but was greater than those in the Central
and  Southern Rocky  Mountain subregions. The
Pacific Northwest had the second highest percentage
of dilute (conductance < 10 fjS cm"1) lakes (26.7%)
and the second lowest Qi value (7.8 /uS cm"1. Table
5-21) for conductance of the western subregions.
Although the Pacific Northwest contained  fewer
dilute and low ANC lakes than did California, many
of these  lakes are comparable to  the most dilute
waters sampled anywhere in the  West.

The Pacific Northwest had the lowest watershed area
to lake area ratios of any subregion. Precipitation
is greater than in other subregions. Adeepsnowpack
is accumulated,  which melts slowly in some areas.

Lakes  in the Pacific Northwest had the highest
subregional  median value for CI" (Table 5-21) and
the second highest median for Na+ (Table 5-1 9). The
relative abundance for  the  major  cations  in the
Pacific Northwest was Ca > Na > Mg > K (Table
6-5). This order  has also been reported by Wissmar
et  al.  (1982) for subalpine lakes  in the Lake
Washington basin and can be observed in data from
Nelson and Delwiche (1983). The concentrations of
Mg+2 in the Pacific Northwest were relatively low
compared to other subregions. Some lakes, however,
can  have high Mg+2, such as those located on dunite
in the North Washington Cascades (Brakke  and
Loranger 1986). The  second highest subregional
percentage of lakes with S04"2 > 50 /ueq L"1 (17.1 %)
occurred in the Pacific Northwest (Table 5-12).

The six geomorphic units for the Pacific Northwest
(Figure 7-2) were delineated on the basis of elevation
and major mountain passes. The Puget Lowlands
included lakes less than 610m (2000 ft) in elevation
in Washington. The North  Washington Cascades
included lakes from the Canadian border south to
Rainy  Pass.  The Middle  Washington Cascades
contain lakes between Rainy Pass and Stevens Pass.
Lakes between Stevens Pass and  Snoqualmie Pass
were  included in the Wenatchee  Mountains.  The
Southern  Washington  Cascades  covered lakes
between Snoqualmie Pass and the  Columbia River
(Oregon border).  Lakes  at elevations above 914 m
(3000 ft) in Oregon were included in  the Oregon
Cascades. These five geomorphic units represented
93 percent of the estimated target population of lakes
in the Pacific Northwest. The Olympic Uplift was also
designated as a geomorphic unit, but the insufficient
sample size  greatly reduced the certainty in
population estimates.

7.2.1  The North Cascades

The North Washington Cascades,  Middle Washing-
ton Cascades,  and  Wenatchee Mountains  are
comprised largely of Paleozoic sedimentary rocks,
which have  been  folded  and metamorphosed,
although  some younger  metamorphic rocks  and
intrusions of granitic rocks are present. Unlike the
North Cascades, the bedrock of the South Washing-
ton Cascades and  Oregon Cascades  is  largely
volcanic.

The  North Cascades  cover an  area of about
20,000 km2 which is almost entirely forested except
for the highest peaks (Franklin and  Dyrness 1973).
Above 1800 m permanent snow fields comprise 13
percent of the area (Rasmussen and Tangborn 1976).
Pleistocene glaciers carved deep U-shaped valleys,
and many active glaciers are still present. Many small
lakes  occur (Qi for lake area = 1.9  ha, Table 7-4).
Bedrock types are  very diverse,  especially in the
North Washington Cascades, where metamorphosed
Paleozoic granitic and  volcanic rocks  occur with
Quaternary volcanic. Cretaceous sedimentary,  and
pre-Jurassic  metamorphic  rocks.  This geological
diversity is reflected in the variability of the chemistry
of lakes.

In the Wenatchee Mountains, the  median ANC was
104.8 /ueq L"1 and an estimated 86.8 percent of the
lakes  had ANC < 200 /ueq L"1. This area contained
the second highest estimated number of lakes with
ANC < 50 /ueq L"1 in Washington.  These  findings
agree with those of Welch et al. (1986) for the Alpine
Lakes  Wilderness  which comprises most of the
Wenatchee Mountains.

An estimated  1 5.0 percent of the lakes in the North
Washington Cascades had ANC < 50 /ueq L"1 (Table
7-3); 49.0 percent had  ANC < 200 //eq L"1. These
percentages compare favorably with results obtained
                    124

-------
Table 7-4.   Population Statistics for Selected Variables for Lakes in Geomorphic Units in the Pacific Northwest (4B),
           Western Lake Survey - Phase I
Conductance
(l*S cm"1)
Mg + 2
(Meq L 1)
K +
cr
Si02
(mg L~1)
Fe
Geomorphic Unit
                    Q-|
                         M
Q,
                                     M
0-1
M
                                                            M
                                                                       M
                                                                                  M
Q, = first quintile (20th percentile).
M = median (50th percentile).*
3These three geomorphic units together form the North Cascades geomorphic unit.
bWatershed  area:lake area.
                                                              M
Oregon Cascades
S. Washington Cascades
North Cascades
Wenatchee Mountains3
M. Washington Cascades3
N. Washington Cascades3
Puget Lowlands



Geomorphic Unit
Oregon Cascades
S. Washington Cascades
North Cascades
Wenatchee Mountains3
M. Washington Cascades3
N. Washington Cascades3
Puget Lowlands
4.4 11.3
7.2 17.6
7.8 14.8
6.5 13.9
9.0 13.6
8.1 27.1
22.3 34.1

Site Depth
(m)
Q! M
2.5 4.8
3.4 7.5
3.9 10.6
3.5 11.1
3.1 8.2
5.4 10.4
2.2 5.5
9.2 23.9
10.9 23.5
8.4 16.7
7.6 13.6
9.7 16.7
6.8 28.7
46.5 79.2

Lake Area
(ha)
Q, M
2.4 3.4
2.0 4.8
1.9 4.2
2.0 4.5
1.8 3.2
1.5 5.9
3.2 6.7
17.6
23.4
13.7
15.6
12.8
11.5
47.6


WA:
QI
5.3
6.7
11.1
11.3
11.1
7.2
9.8
32.6
34.5
19.3
21.4
18.5
19.9
67.4


:LAb
M
21.4
12.6
19.5
20.0
19.7
12.0
22.8
2.6 4.6
2.0 3.8
1.9 4.5
1.6 4.1
2.9 8.7
1.5 3.4
4.3 6.2
Secchi Disk
Transparency
(m)
Q! M
2.3 4.7
2.7 4.6
3.5 7.5
3.5 8.1
2.7 5.8
4.6 7.9
1.8 2.5
6.0
11.6
2.5
3.7
2.4
1.7
24.4











8.2
14.5
4.7
9.1
3.5
2.4
38.2











1.8
1.1
1.6
1.6
1.3
1.6
1.7











4.5
3.0
2.7
2.7
2.7
2.3
5.0











1.9
8.0
3.8
5.1
0.0
2.6
21.6











10.6
20.8
12.4
10.2
13.5
13.4
58.5











 by Brakke and Loranger (1986) from a survey of 57
 lakes in the same area. They calculated that 15.8
 percent of the lakes had ANC < 50 (jeq L~1 and 66.7
 percent had ANC  < 200 /jeq L"1.  Although some
 extremely low ANC values (3-5 /jeq L~1) have been
 observed in the North Washington Cascades, ANC
 has  been  reported  to  be  variable  and strongly
 associated with  bedrock  geology  (Brakke and
 Waddell  1985; Brakke  and Loranger 1986). The
 interquintile difference (Qd) for ANC in the present
 study was  320 peq L"1  for the  North Washington
 Cascades. Lakes in the Middle Washington Cascades
 and  Wenatchee  Mountains  showed much  less
 variability in  ANC (Qd  =  128 and  107  (jeq  L'1,
 respectively).

 As  in  the case  of the West in general,  Ca+2
 concentrations in North Cascade lakes were similar
 to ANC concentrations. In  watersheds within  the
 Cascades, weathering rates are high for some lakes,
 particularly those  most  recently deglaciated (Rey-
 nolds and Johnson  1972). Relatively small areal
 exposures of calcite  can generate significant
 alkalinity and Ca+2 (Drever and Hurcomb 1986) as
 a result  of weathering. However, when  calcite
 materials are absent, very dilute lakes with low ANC
can result.

 Sulfate concentrations in the Middle and North
Washington Cascades were the highest found in the
 Cascade range (M = 23.8/jeq  L"1 and M = 21.6 (jeq L"1,
                                                  respectively). The lowest values of SCu 2 in Washing-
                                                  ton were found in the Wenatchee Mountains (Qi
                                                  = 5.4 fjeq L"1; M =  12.6 yueq L"1, Table 7-3).

                                                  Dissolved organic carbon concentrations were very
                                                  low in the North Cascades. The estimated Q4 value
                                                  was 1.43 mg L"1 and 90.4 percent of the lakes had
                                                  DOC < 2 mg L"1 (Table  7-3). Similarly, conductance
                                                  was low, with 20 percent of the lakes having values
                                                  < 7.8 fjS cm~1 (Table 7-4).

                                                  7.2.2  The  Oregon Cascades

                                                  The Oregon Cascades are more uniform geologically
                                                  than  the  Washington  Cascades,  although two
                                                  separate units can be recognized. These are the High
                                                  Cascades  on the  East and  the  West  Cascades,
                                                  running parallel from  north  to south. The West
                                                  Cascades are older and lower in elevation than the
                                                  High Cascades (Baldwin 1981). However, interme-
                                                  diate volcanic rocks occur in both areas and because
                                                  of the uniformity of chemistry in lakes, we have not
                                                  subdivided the Oregon Cascades.

                                                  The  Oregon Cascades have the  second  lowest
                                                  median ANC (92.1 //eq  L'1) and the second highest
                                                  percentage of lakes with low ANC (25.5% with ANC
                                                  < 50  fjeq L"1) in the Pacific Northwest (Table 7-3).
                                                  These estimates agree  with the median ANC (83.9
                                                  /jeq L"1) for  63 Oregon Cascade lakes sampled by
                                                  Nelson and  Delwiche (1983). Calcium concentra-
                                                                          725

-------
tions in the Oregon lakes are among the lowest of
all geomorphic units, with a median of 48.6 //eq L~1
and 55.8 percent of the lakes having Ca+2 < 50
//eq L"1 (Table 7-3). Similar results were obtained
by Nelson and Delwiche (1983); the mean Ca+2 was
57.9 //eq L~1. Lakes in the Oregon Cascades showed
relatively little interquintile difference (Qd) in Ca+2
andAIMC.

Lakewater  S04~2 concentrations in  the Oregon
Cascades and in the adjacent California  Cascades
were the lowest in  the West, and are among the
lowest observed in the  Northern Hemisphere
(Armstrong  and Schindler 1971). Lakes  in the
Oregon Cascades had the lowest median (1.6//eq L"1)
and the second lowest 04 (4.0 //eq L~1) value for
SO*"2 of  any geomorphic  unit (Table 7-3).  These
results are  comparable  to those of  Nelson  and
Delwiche (1983), who found a mean SO4~2 concen-
tration of 5.82 //eq L"1 with a minimum concentration
of 0.62//eq L"1.

Lakes in the Oregon Cascades are  generally shallow
(median site  depth  = 4.8  m) and dilute (median
conductance = 11.3 //S cm"1. Table 7-4). It appears
that Na+ and CI" are generated from watersheds in
both the Oregon Cascades and the South Washing-
ton Cascades. Even though more distant from marine
waters, the Q, and median values of Na+ and CI"
were  much greater than  in  lakes in  the  North
Cascades (Table 7-4). Similarly, Nelson and Delwiche
(1983) observed moderate concentrations of Na+ and
CI" in the Oregon Cascades. Nelson (1 985) analyzed
data from Oregon Cascade lakes and concluded that
lakewaters  were in equilibrium with kaolinite and
gibbsite.  The slow dissolution  kinetics  of  these
minerals  were reflected in very  dilute lakes. The
weathering products of feldspars  include Na+ and
K+, which contribute to lakewater  concentrations of
these cations.

7.2.3  Other Geomorphic Units

In contrast to the North Cascades,  the  South
Washington  Cascades contains   few  areas with
exposed sedimentary  or metamorphic rocks. The
South  Washington  Cascade Range is  comprised
mainly of andesite and basalt and is heavily forested.
This geomorphic unit is  much more uniform than
the North Cascades, even though it includes a series
of recent Quaternary deposits related to the volcanic
activity of Mount Rainier, Mount Adams, and  Mount
St. Helens.

The  lowest Qi  for ANC  (21.3  //eq  L"1) of any
geomorphic unit (Table 7-3) was  estimated for the
South Washington Cascades. In the Pacific Northw-
est, the highest percentages of lakes with low Ca+2
and low  ANC occurred  in the South Washington
Cascades, but lakes  in this geomorphic unit  also
showed the most variability in these variables (Qd
for ANC = 355 //eq L"1, Qd for Ca+2 = 246 //eq  L"1).
Relatively little sampling has been conducted in the
South Washington Cascades, although some  low
ANC lakes are known to occur (Duncan and Ausserer
1984; Wissmar et al. 1982; Bortelson et al. 1977).
Concentrations of DOC were slightly higher in the
South  Washington  Cascades  than in  the North
Cascades, although none of the lakes sampled had
DOO6 mg L"1.

Concentrations of all measured constituents were
relatively high in lakes in the Puget Lowlands (Tables
7-3 and 7-4). A low  percentage (8.9%)  of lakes  in
the Puget Lowlands had ANC < 50 //eq L"1 (Table
7-3). As expected from the proximity of these lakes
to coastal waters, the highest concentrations of Na+
and CI" of any geomorphic unit occurred in the lakes
of the Puget Lowlands. The Puget Lowlands had the
highest percentage (53.8%) of lakes with S04~2 > 50
//eq L"1 and the highest median value for S04~2 (50.2
//eq L~1) of any geomorphic unit (Table 7-3).

Dissolved organic carbon concentrations in the Puget
Lowlands were very high for the West, with the Qi
value (2.26 mg L"1) exceeding the highest Q4 value
in any geomorphic unit of the Cascade Range (Table
7-3). These  high  DOC concentrations  may be
attributable to greater production of organic matter
in lakes  related to higher concentrations of total
phosphorus (Sumioka and  Dion 1985; Gilliom  and
Bortelson 1983).
7.2.4  Sulfate in Geomorphic Units in the Pacific
Northwest

Sulfate concentrations in lakes increased north  of
the Columbia  River, and  were markedly higher  in
Washington than in Oregon. The estimated Qi value
for SO4~2 (10.8 //eq L"1) in  the South  Washington
Cascades is more than twice the Q4 value for S04~2
in the Oregon Cascades (4.0 //eq  L"1). Along the
Cascades, the concentrations of  SO4~2 are  low  in
California and  highest  in the North Cascades. Within
the North Cascades, lakewater S04~2 concentrations
are highest in  the North Washington Cascades and
lowest in the Wenatchee Mountains.

Sulfate can be derived from the weathering of sulfur-
containing minerals within watersheds. In the North
Cascades, Loranger (1986)  reported that the
watershed source of S04~2 exceeds the  atmospheric
contribution in some  lakes, indicating weathering
from bedrock as the likely source. Geological sources
of SO4~2 are also indicated for some  lakes in the
Alpine Lakes Wilderness (Welch et al. 1 986). In these
cases, S04"2 can be a major anion resulting  in a
                     126

-------
small deficit of ANC relative to Ca^ and  Mg+2
(Loranger 1986; Welch et al. 1 986).

Some of the SO/2 could also be derived from sea
salts  contained in precipitation. To  examine the
contribution of  sea salt to lakewater S04 2 concen-
trations, SO/2 was corrected for sea salt using the
formula in Gorham et al. (1985). The results, which
are not  shown, indicated that very little of the lake
SO/2 is derived from sea salt.

Sulfur  contained in emissions from volcanic
eruptions can  serve as  sources of SO/2 to lakes.
The recent eruption of Mount St. Helens temporarily
may have increased SO/2 concentrations in  lakes
in the surrounding area;  Wissmar et al. (1982)
reported changes in lakewater chemistry near the
blast zone. However, if the lakes in this area have
short hydraulic residence times, as calculated on the
basis of map information (Section 6.4.3.1), and hence
rapid flushing  rates, it is likely that the Mount  St.
Helens  eruption had only a short-term effect  on
lakewater SO/2 concentrations for most lakes in the
area.
7.3   Northern Rocky Mountains (4C)

The  Northern  Rocky Mountains (Subregion 4C) is
geologically distinct from the Central and Southern
Rockies. The chemical characteristics of the lakes
in this subregion are also distinct from those of lakes
in the Central  (4D) and  Southern  (4E) Rocky
Mountains. Geological and, perhaps, meteorological
differences likely account for this variability. Lakes
in the Northern Rockies were lower in elevation (M
= 2064 m. Table 5-6) than in the Central  Rockies
(M =  3040 m) and  in the Southern Rockies (M =
3264 m).  Only lakes in the Pacific  Northwest (4B)
were lower in elevation (M = 1437) than those in
the Northern Rocky Mountains.

About one-half (50.7%) of the estimated number of
lakes in the target population in the Northern Rockies
(N = 2379) were estimated to have ANC < 200 fjeq L~1
(Table 5-9); in the West, only the Southern  Rockies
had  a lower percentage of lakes in this category
(39.4%).  The median ANC  values  in these  two
subregions were the highest in  the West at 195.2
fjeq L~1 for the Northern Rockies and 317.0 fjeq L~1
for the Southern Rockies (Table 5-16).

As in other western subregions,  pH values for lakes
in the Northern Rockies were generally circumneu-
tral. The Qi value for pH, for example, was 6.90 and
for other subregions ranged from 6.59 in California
to 7.10 in the Southern  Rockies. The median pH
(7.35) in the Northern Rocky  Mountain lakes  was
the second highest observed in the West behind the
Southern Rockies (7.60, Table 5-16).
Sulfate concentrations in lakes of the Northern Rocky
Mountains were generally low (M = 15.6 fjeq L"1,
Table 5-16) and were intermediate when compared
to the other western subregions.  Two other subre-
gions had greater SO/2 medians—the Central and
Southern  Rockies with concentrations of 24.4 and
34.6 fjeq L"1, respectively. However, at the Q4 level,
SO/2 was 32.3 fjeq  L"1 in the Northern Rockies and
only California had a lower Q4 value for SO/2 (13.5
fjeq I/1).

Lakes  in  the  Northern Rockies  had  the  second
highest median  conductance  in  the  West  (25.1
/jS cm"1) and the highest Q4 value (108.9 fjS  crrf1;
Table 5-21). Calcium  values for  the  lakes in this
subregion were also high (M = 154.2 fjeq  L~1 and
Q4 = 730.0 fjeq L"1)  (Table 5-16).  Dissolved organic
carbon was generally low in western lakes  and the
Northern Rockies had the second  lowest population
estimates for this variable; 70.8 percent of the lakes
in this subregion had DOC < 2.0 mg L"1. There were
few  lakes with large surface area in the Northern
Rockies (M =  4.6 ha  and  Q4 = 8.9 ha). Only the
Southern Rockies had smaller lakes.

Nine geomorphic units in the Northern Rockies were
delineated based on mountain  ranges (Figure 7-3).
The  Clearwater Mountains, Salmon  River Moun-
tains, Sawtooth Mountains, and Bitterroot Range are
collectively identified and discussed in this section
as the Idaho  Batholith. A total of 88 lakes  were
sampled in the Idaho Batholith, 37 of which are  in
the Bitterroot Mountains. Because many lakes were
sampled in the Bitterroots, this geomorphic unit  is
also discussed separately. Population  estimates
were also calculated for the Lewis Range, the Blue
River Uplift, and the Anaconda-Pintlar Mountains,
each of which includes several distinct mountain
ranges. However, because relatively few lakes were
sampled in these latter geomorphic units, population
estimates have large upper confidence limits (Ncu,
Section 7.1.2). The geomorphic  units for which
population estimates were calculated represent 72
percent of the target  population  in the Northern
Rockies.  The  Selkirk  Mountains  and the  Cabinet
Mountains were  also  designated  as  geomorphic
units, but due to the small number of lakes sampled,
population estimates are not presented.

7.3.1  Idaho Batholith

The  Northern Rockies  is largely formed of  granitic
intrusions of Cretaceous age.  This region  is
dominated by the  Idaho  Batholith,  a  roughly
rectangular area 140  km wide and 390 km long
which ranges in elevation from about 900 to 2100  m
(Hunt  1974).  Approximately  85 percent of the
batholith,  which is wholly north of the Snake  River
in Idaho,  is composed  of granitic rocks;  related
                                                                         727

-------
Figure  7-3.    Locations of sampled lakes within geomorphic units in the Northern Rockies (4C), Western Lake Survey - Phase I.
              Lakes located outside the boundaries were not included in estimates for these subpopulations.
                                        Selkirk
                                        Mountains
                                                            4C
Cabinet
Mountains
                                                                          Bitterroot
                                                                          Range
                                                                                Lewis Range
                                                                                Anaconda-Pintlar
                                                                                Mountains
                                                                                              r
                                                   Sawtooth
                                                   Mountains
                                               ID
                                                                                  	-1
                         128

-------
portions  of the batholith are  located in  Montana,
Oregon, Washington and British Columbia  (Larson
and Schmidt 1958).

Median ANC for the Idaho Batholith is 98.3 /ueq L~1
(Table 7-5). The Qi  and Q4 values for ANC in this
geomorphic  unit  were  43.8  and  396.0 /ueq L"1,
respectively.  In comparison,  lakes  of the Sierra
Nevada had a median ANC  of 59.8 /ueq L~1 and Qi
and Q4 of 33.6 and 126.2 peq L"1, respectively (Table
7-1). Conductance was higher in the Idaho Batholith
(M =  11.5 /uS  crrf1,  Table 7-6) than  that observed
in the Sierra Nevada (M = 8.2 jjS crrf1. Table 7-2).

7.3.2  Bitterroot Range

The Bitterroot Mountains lie  on the northeastern
edge of the Idaho Batholith, an area dominated by
igneous rocks of Mesozoic age. An estimated 77.6
percent (Nc = 220) of the lakes in this mountain range
had ANC < 200 peq L"1 and 33.1 percent had ANC
< 50 fjeq L"1  (Table  7-5). The  Bitterroot Mountains
contained the third lowest median ANC value of any
geomorphic unit in the  West  (70.4 /ueq L"1). Only
lakes  in  the  Sierra  Nevada and  in  the  Sawtooth
Mountains  had lower median ANC values  (59.8
/ueq L~1 and 67.8 /ueq  L"1, respectively) (Table 7-1).
Lakes in the Bitterroots also had the  second lowest
median Ca+2 value (48.3 /ueq L"1). Conductance for
lakes in the Bitterroots was very low (M = 9.8 fiS cm"1,
Table  7-6); only in the Sierra Nevada was conduc-
tance  lower (M =  8.2  pS cm"1. Table 7-2).  Sulfate
was also low  in the  Bitterroot Mountains (M  = 9.0
/aeq L~1) and it was estimated that 80 percent of the
lakes had SO4~2 concentrations less than 17.1 /ueq
L"1 (Table 7-5).

7.3.3  Other Geomorphic Units

Of the geomorphic units in the Northern Rockies,
the Lewis Range had the largest estimated number
of lakes in the target population (N = 553). This area
contains Glacier National Park and is dominated by
Precambrian  sedimentary  rock which  has  been
metamorphosed,  largely, into quartzite  (Alt  and
Hyndman 1978). Only 15 lakes were sampled in this
geomorphic unit;  therefore, the upper  confidence
limit (Ncu) on  statistical estimates  is  large.  Recog-
nizing  this limitation, lakes  in  this area were
estimated to have high ANC relative to lakes in other
geomorphic units  (M = 392.2  /ueq  L"1; Table 7-5).
The percentage of lakes with  ANC < 200 fieq  L"1
in this area (33.3%) was the lowest of all geomorphic
units analyzed in the Northern  Rockies.
Only 12 and 10 lakes, respectively, were sampled
in  the Anaconda-Pintlar Mountains  and the Blue
River Uplift. As in for the Lewis Range,  population
estimates  for these geomorphic  units have large
upper confidence limits (Section 7.1.2). No lakes with
ANC < 50 /ueq L"1 were sampled in either of these
areas, but their median ANC values were less than
that estimated for the Lewis Range (Table 7-5). The
Anaconda-Pintlar Mountain lakes had the highest
median S04"2 concentration in the Northern Rockies
(32.0 /ueq L"1).

7.4  Central Rocky Mountains (4D)

Although less than  seven percent  of the lakes in
the Central Rockies  (Subregion 4D) had ANC < 50
/ueq L"1, 77.8 percent had ANC < 200 /ueq L"1 (Table
5-9), which was the second highest percentage in
the West next to California. The Central Rockies had
the second lowest median ANC (104.9 //eq L~1, Table
5-16). Lakes  in the Central Rockies also had the
second highest Qi values, but the second lowest M
and Q4 values for Ca+2 and conductance (Tables 5-16
and  5-21). Extractable aluminum  concentrations
were the lowest among all subregions at the Qi,
M, and Q4 levels.

The relative abundance of major cations at both Qi
and median values was Ca > Mg > Na  > K (Table
6-5). The median Na+ concentration was the second
lowest among western subregions. Nearly all of the
lakes in the Central Rockies were clearwater (true
color < 30 PCU) with Ca+2 and HC03" as the dominant
ions (Section 6.2.6). However, in some lakes Na+ was
the dominant  cation and  HCOs" or S04"2 was the
dominant anion. The Central Rockies had the lowest
percentage of lakes with S04"2 > 50 /ueq L"1 among
western subregions.

An estimated 235 lakes (10%) in the Central Rockies
had conductance <  10 /uS cm"1, the second lowest
percentage of dilute lakes among western subre-
gions. Although evaporation rates in this subregion
are comparable  to  those observed for lakes in
California and in the Central Rockies (Kohler et al.
1955),  total precipitation  is  less  in the Central
Rockies  (U.S.  Department of Commerce 1974). As
a consequence, higher concentrations  of solutes
were observed for lakes in the Central Rockies. Lakes
in this subregion also may receive atmospheric
contributions  of materials  in the form  of airborne
dust containing gypsum or calcium carbonate from
the surrounding  arid basins.  This influence is
indicated by higher  Ca+2 concentrations in precip-
itation measured at NADP/NTN sites in Colorado and
Wyoming (15  /ueq L"1) than those measured at sites
in California,  Oregon,  and Washington (3 /ueq L"1,
Hidy and Young 1986). Hidy and Young (1986)  also
found that Ca+2 explained 85 percent of the variance
in S04"2 concentrations, which  supports  the
hypothesis that wind-borne alkaline dust comprised
of CaCOs and CaS04 influences lake chemistry in
this subregion.
                                                                         129

-------
Table 7-5.    Population Estimates and Statistics for Primary Variables by Geomorphic Unit in the Northern Rockies (4C),
             Western Lake Survey - Phase I
ANC;
Geomorphic Unit n*" N pc
Lewis Range 15 553 0.067
Anaconda-Pmtlar Mtns. 12 239 0.000
Blue River Uplift 10 73 0.000
Idaho Batholith 88 937 0.236
Bitterroot Range3 37 283 0.331
Salmon River Mountains3 22 379 0.127
Sawtooth Mountains3 20 192 0.228
SO4"2
Geomorphic Unit n*** N pc
Lewis Range 15 553 0.067
Anaconda-Pintlar Mtns. 12 239 0.356
Blue River Uplift 10 73 0.157
Idaho Batholith 88 937 0.000
Bitterroot Range3 37 283 0.000
Salmon River Mountains3 22 379 0.000
Sawtooth Mountains3 20 192 0.000
~ 50 (jeq
NC
37
0
0
221
94
48
44
> 50 /jeq
Nc
37
85
12
0
0
0
0
L-1
NCU
97
—
—
297
125
111
67
L"
NCU
97
171
30
—
—
—
-
ANC
QI
77.2
143.7
95.9
43.8
38.9
71.9
44.2
S04-2
QI
8.0
18.5
4.3
5.5
5.4
6.3
5.9
(MeqL
M
392.2
255.0
123.4
98.3
70.4
177.4
67.8
(/^eq L
M
10.3
32.0
5.7
8.8
9.0
10.7
7.5
-1) Ca*2 < 50 yueq L~1
Q4
1208.3
469.3
178.5
396.0
285.8
421.3
884.0
-1)
Q4
32.2
58.6
23.0
17.4
17.1
15.7
20.8
PC
0.200
0.048
0.171
0.344
0.500
0.188
0.260
DOC <
PC
0.733
0.596
0.843
0.781
0.809
0.715
0.808
Nc
111
11
12
322
142
71
50
2 mg
Nc
406
142
61
731
229
271
155
NCU
213
30
26
406
184
138
75
I'1
NCU
587
236
92
859
288
389
209
Ca4
Q,
49.0
148.7
57.6
31.2
26.8
50.7
40.8
DOC
QI
0.35
0.92
0.37
0.81
0.79
0.99
0.38
-2 / i — 1 \
M
328.0
202.0
87.2
63.7
48.3
72.0
63.5
(mg L~
M
0.78
1.67
0.70
1.22
1.23
1.54
0.87
Q4
832.6
326.0
131.3
360.2
216.4
368.8
805.6
1,
Q4
2.67
2.28
1.59
2.08
1.93
2.30
1.82
n*" = number of lakes from which samples were obtained.
N = estimated number of lakes.

pc = estimated proportion of lakes meeting the given
Nc = estimated number of lakes meeting the
Ncu = 95 percent upper confidence limit for N
Q, = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
( — ) = undefined.

criteria


















given criteria.
C"






















































aThese three geomorphic units, together with the Clearwater Mountains (not presented separately), form the Idaho Batholith
 geomorphic unit.
Table 7-6.    Population Statistics for Selected Variables for Lakes in Geomorphic Units in the Northern Rockies (4C),
             Western Lake Survey - Phase I
Conductance Mg + 2 NE
(^iS cm~1) (Mecl L ) (^eP
Geomorphic Unit
Lewis Range
Anaconda-Pmtlar Mtns.
Blue River Uplift
Idaho Batholith
Bitterroot Range3
Salmon River Mountains3
Sawtooth Mountains3


Q!
9.0
18.4
11.2
6.6
5.4
8.9
7.7

Site
M
39.1
32.6
13.1
11.5
9.8
18.4
10.5

Depth
(m)
Geomorphic Unit
Lewis Range
Anaconda-Pmtlar Mtns.
Blue River Uplift
Idaho Batholith
Bitterroot Range3
Salmon River Mountains3
Sawtooth Mountains3
QI
3.1
3.3
1.4
4.6
5.4
4.8
4.8
M
11.0
5.4
3.4
10.1
11.8
8.6
9.9
QI
30.4
15.6
7.9
6.6
5.5
9.1
5.4

Lake
M
110.6
22.5
9.2
10.4
13.2
17.4
8.4

Area
(ha)
QI
2.3
4.2
2.1
2.2
2.6
2.1
2.4
M
4.9
4.8
2.9
3.7
4.8
2.5
5.4
QI
4.0
28.5
20.8
13.5
11.9
21.9
16.7


WA;
Q,
6.0
10.6
20.4
9.9
10.3
10.5
5.7
L-1)
M
9.7
43.1
26.0
23.1
15.7
38.8
22.6


:LAb
M
13.6
23.8
23.0
18.5
16.0
20.8
20.4
K +
Q, M
2.1 2.8
3.2 6.0
9.1 10.4
2.9 4.7
2.6 4.1
4.4 5.8
2.6 3.0
Secchi Disk
Transparency
(m)
Q, M
2.0 4.8
1.5 4.0
1.4 3.5
3.7 6.3
4.1 6.4
2.4 6.3
4.2 5.8
Cl
(^eq I
QI
1.6
2.7
1.4
1.7
1.8
1.7
1.7











SiO2
."'I (mg L-1)
M Q,
2.1 0.8
3.5 3.0
1.9 2.2
2.4 2.1
2.3 1.7
2.4 2.6
2.4 2.7











M
1.5
4.2
3.0
3.3
2.2
3.8
3.9











Fe
(/"9 I-'1)
Q, M
3.0 7.8
9.7 15.8
5.3 14.0
3.7 9.7
5.6 11.5
4.1 7.4
0.0 8.0











QT  = first quintile (20th percentile).
M  = median (50th percentile).
"These three geomorphic units, together with the Clearwater Mountains (not presented
 separately), form the Idaho Batholith geomorphic unit.
bWatershed area:lake area.

                           730

-------
The median lake elevation (3040 m) in the Central
Rockies was  the  second highest among  western
subregions. Median values for watershed area (178
ha), lake area (5.4 ha), and watershed area to lake
area ratio (30.2) were the highest among  western
subregions (Table 5-6); however, watershed area in
the Central Rockies was also the most variable (Qd
 551 ha).
The  eight geomorphic units in the Central  Rockies
(Figure 7-4) were delineated on the basis of major
mountain ranges. Population estimates and statistics
were calculated for the Beartooth Uplift, the
Yellowstone Plateau, the Bighorn Uplift, the  Wind
River Uplift,  and the  Uinta Arch.  These  five
geomorphic units represented 90 percent of the
target lake population in the  Central  Rockies (N =
2299). The Teton Uplift, the Gros Ventre Uplift, and
the Wasatch Mountains  were also designated as
geomorphic units;  however,  fewer than 10  lakes
were sampled  in each of these areas, so population
estimates are not presented.

7.4.1  The Wind River Uplift

The  Wind River  Range is an uplift of Precambrian
granites and gneisses. Glacial cirques are numerous
and  many ponds less than 1 ha in area are present.
The  lakes in the estimated target population (N =
884) were larger than those in any other geomorphic
unit; they  were also relatively  deep  (Table  7-8).
Several  active  glaciers remain in the alpine tundra
above 3350 m, particularly on the east side of the
Continental Divide (Arno 1969). The southern and
eastern  portions receive only  about  50 cm of
precipitation,  whereas the northwestern  portion
receives more precipitation (Steele et al. 1 983).

The  median ANC value for lakes  in the Wind  River
Range was 109.3 /c/eq L~1  (Table 7-7).  This was
slightly higher than the means  of 90 /ueq  L"1 and
60 /ueq L"1 measured in lakes on the  western and
eastern  slopes,  respectively,  in  a  Forest  Service
survey of 92 lakes in the Wind River (Stuart 1984,
Galbraith 1984).

The  Qi value for SO-f2  in Wind River lakes  (21.7
yueq L"1) was among the highest for all geomorphic
units; the median value was 24.3 fjeq L~1 (Table 7-7).
There was little interquintile difference in S04"2, with
a Q4 value of 31.3 /ueq L~1. These values are very
similar to those  measured in  a study by Galbraith
(1985) in the Wind Rivers in which the reported range
was  20-35 //eq L"1.

The  Qi value for conductance was 12.4 /jS cm'1,
which was close to the median value of 15.0/uS cm"1
(Table 7-8). As in the other geomorphic units in the
Central  Rockies, relatively few lakes  had conduc-
tance < 10/uS cm  , but many lakes had conductance
between 10-1 5 /uS cm"1.

7.4.2  The Uinta Arch

The Uinta Arch is a large uplift running east to west.
Glacial cirques and lakes are numerous at  higher
elevations.  The  mountains consist  mainly of
Precambrian sedimentary and metamorphic rocks,
particularly quartzites. At  lower elevations, lime-
stones, sandstones  and  shales  are common.
Timberline occurs at about  3350  m  (Arno  and
Hammerly 1984). Lakes in  the Uintas are smaller
and shallower than lakes in the rest of the Central
Rockies (Table 7-8).

Similar to the Bighorn and Wind River areas, the
Uintas  have a high percentage of lakes with ANC
< 200 /jeq L"1 (93.1 %) but a  low percentage of lakes
with ANC < 50 /ueq L"1 (3.9%).  Results for Ca+2 were
similar (Table 7-7). The  Qi  and  median values for
conductance were 11.5 and 15.1 juS cm"1, respec-
tively, which were nearly identical to those for lakes
in the Wind  River   Range.  Chloride and iron
concentrations were  higher in lakes in  the  Uintas
than in other geomorphic units in the Central
Rockies. The  median SiO2 concentration  in the
Uintas  (1.0 mg L"1) was the  lowest estimated for
any geomorphic unit (Table 7-8).

7.4.3  Other Geomorphic Units

The Beartooth Uplift contains many lakes, with about
80 percent at elevations > 2700 m. The Beartooth
Uplift  is a large  (5200 km2),  rectangular  block of
Precambrian  igneous  and  metamorphic  rocks
uplifted along faults. The area was originally covered
by sedimentary rocks that subsequently have been
removed by glaciation and other erosive forces (Alt
and Hyndman 1 972). The Beartooth Uplift was once
covered almost completely by a glacier. Lakes in the
Beartooths are deep (M = 11.8 m) and relatively large
(M = 6.3 ha. Table 7-8).

All lakes sampled in the Beartooth Uplift had ANC
< 200 /yeq L"1, and 15.7 percent (the highest among
geomorphic units  in this subregion) had ANC < 50
/ueq L"1 (Table 7-7). However, because only 12 lakes
were sampled in  this geomorphic unit, statistical
confidence in these population estimates is low. In
contrast, the Qi value for ANC (61.3 /ueq L~1)  in this
area was the second highest estimated for geomor-
phic units  in the  Central Rockies. The median Ca+2
value in the Beartooths (69.7/ueq L~1) was the second
lowest  among  the geomorphic units  in the Central
Rockies. Lakes in the Beartooth Uplift had the  lowest
Q4 value for DOC (0.84 mg L"1) among all geomorphic
units. All the lakes sampled had DOC < 2.0 mg L"1,
and the median  DOC value (0.60 mg L'1)  in the
Beartooths was as low as the Qi values  in other
                                                                         737

-------
Figure 7-4.    Locations of sampled lakes within geomorphic units in the Central Rockies (4D), Western Lake Survey - Phase I.
              Lakes located outside the boundaries were not included in estimates for these subpopulations.
                                                                            4D
                     ID
                                   Teton
                                   Uplift
                                                              Beartooth
                                                              Uplift
                              Wasatch
                              Mountains
MT
                                                                                     Bighorn
                                                                                     Uplift
                                     Cros
                                     Ventre
                                     Uplift
                                                       Uinta
                                                       Arch
                                                                                                          ...J
                          UT
                         132

-------
Table 7-7. Population Estimates and
Statistics for Primary Variables by Geomorphic
Unit in the Central Rockies (4D).
Western Lake Survey - Phase 1

Geomorphic Unit n"* N
Beartooth Uplift 12 212
Yellowstone Plateau 1 5 233
Bighorn Uplift 1 1 203
Wind River Range 47 884
Uinta Arch 27 503

Geomorphic Unit n*** N
Beartooth Uplift 12 212
Yellowstone Plateau 15 233
Bighorn Uplift 1 1 203
Wind River Range 47 884
Uinta Arch 27 503
ANC<5(
PC
0.157
0.128
0.097
0.062
0.039
SO/2 >
PC
0.000
0.192
0.000
0.000
0.036
0 fjeq L"1
Nc Ncu
33 71
30 63
20 51
55 105
20 51
50 ,ueq L~1
Nc Ncu
0 -
45 85
0
0
18 48
ANC (neq
0-1
61.3
58.5
52.7
74.0
59.3
S04
QI
13.7
8.2
18.1
21.7
19.5
M
92.3
175.1
82.4
109.3
83.4
^ (Meq
M
20.8
12.6
26.2
24.3
23.7
I'1)
Q4
155.6
705.7
145.4
167.6
112.8
U-1,
Q4
33.6
33.3
35.4
31.3
32.9
Ca+2
PC
0.157
0.212
0.097
0.062
0.000
DOC
PC
1.000
0.617
0.755
0.646
0.524
<50/ueq
Nc
33
50
20
55
0
< 2 mg L
Nc
212
144
153
571
264
LM Ca+2 (Meq
NCU
71
96
51
105
-
-1
Ncu
308
218
235
693
369
QI
48.8
43.8
54.6
70.5
65.3
DOC
QI
0.50
0.80
0.51
0.58
0.79
M
69.7
137.0
67.7
93.8
81.6
(mg L
M
0.60
1.63
0.94
1.17
1.79
L-')
Q4
117.6
426.0
115.3
140.7
93.8
.-,,
Q4
0.84
5.42
2.48
3.19
3.37
n*" = number of lakes from which samples were obtained.
N = estimated number of lakes.

pc = estimated proportion of lakes meeting the

given criteria


















f\lc = estimated number of lakes meeting the given criteria.
Ncu = 95 percent upper confidence limit for Nc.
QT = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
( — ) = undefined.












































 Table 7-8.   Population Statistics for Selected Variables for Lakes in Geomorphic Units in the Central Rockies (4D), Western
            Lake Survey - Phase I
Geomorphic Unit
Beartooth Uplift
Yellowstone Plateau
Bighorn Uplift
Wind River Range
Uinta Arch
Geomorphic Unit
Beartooth Uplift
Yellowstone Plateau
Bighorn Uplift
Wind River Range
Uinta Arch
Conductance
(^S cm"1)
Q! M
8.9 12.0
9.2 23.4
8.3 11.5
12.4 15.0
11.5 15.1
Site Depth
(m)
Q! M
6.3 11.8
4.2 6.1
5.5 8.6
6.6 10.9
3.2 5.7
Mg+2 Na+
Q, M
20.4 37.0
18.3 50.1
14.1 27.5
15.0 22.2
26.1 34.7
Lake Area
(ha)
Q! M
2.9 6.3
2.8 4.3
2.7 4.7
4.0 6.4
2.3 3.4
0-1
8.3
15.3
8.7
15.4
11.2
WA:
QI
28.6
8.0
17.4
13.1
11.7
M
12.0
31.4
16.5
22.9
18.7
:LAa
M
38.6
23.5
41.6
22.5
30.1
K+ Cr SiO2

-------
The lakes with the maximum (4948.6 /jeq L 1) and
the minimum (-24.0 (jeq L"1) ANC measured in the
WLS-I occurred in the Yellowstone Plateau. Of all
lakes sampled in the geomorphic units in the Central
Rockies, those in  the Yellowstone  Plateau had the
highest median and Q4 values for ANC, pH, and Ca+2.
However, this area was the only one containing lakes
with pH <  5.0,  and it had the  highest percentage
of lakes with Ca+2 < 50 fjeq L"1 (21.2%, Table 7-7)
in the Central Rockies. The Yellowstone Plateau also
had the second highest percentage of lakes with ANC
< 50 /jeq  L"1 (12.8%,  Table  7-7). This observed
variability in the population estimates for ANC, pH
and Ca+2 can be explained  in part by the variability
in geology and the variety  of rock  types subject to
mineral weathering (Alt and Hyndman 1972).

First quintile and  median  S04"2 concentrations in
Yellowstone lakes were the lowest in the  Central
Rockies (8.2 and  12.6 /jeq L~1, respectively.  Table
7-7); however, the Yellowstone Plateau contained
the highest percentage of lakes with  S04~2  > 50
/jeq L"1.  Some of the  lakes  on the  Yellowstone
Plateau are probably influenced by natural emissions
of sulfur or contact with sulfate-rich groundwater
as evidenced by the extremely high S04~2 concen-
trations in lakes adjacent to active hot springs. The
Yellowstone Plateau had the highest percentage of
lakes (19.2%) with DOC > 6  mg L"1 of any geomorphic
unit in the WLS-I.

The Bighorns are  anticlinal uplifted, with Precam-
brian bedrock that rises between the arid Bighorn
and Powder River basins.  The granitic core of the
uplifted  dome,  which forms a broad plateau, is
surficially exposed. The mountains are relatively dry,
receiving less than 63.5 cm of precipitation annually
(Hoffman and Alexander 1976). Several areas are
above timberline,  which is  located at approximately
2900 m (Arno and Hammerly 1984). Eleven  lakes
were sampled in  the Bighorns. Within the  Central
Rockies,  lakes in  the Bighorns had the lowest Qi
value for ANC (52.7 fjeq L"1), but only 9.7  percent
of the lakes in the Bighorns had ANC < 50 //eq L"1
(Table 7-7). Sulfate  concentrations were  slightly
higher in  lakes  in the Bighorns than in  other
geomorphic units  in the Central Rockies. No lakes
sampled in  the Bighorn Uplift had DOC > 6 mg L~1.

7.5  Southern Rocky Mountains (4E)

The Southern Rocky  Mountains (Subregion 4E) is
a continuous high-mountain barrier extending  for
about 725 km from Laramie, Wyoming in the  north
to Santa Fe, New Mexico in the  south.  These
mountains are the highest  in the Rocky Mountains.
All  mountain ranges here have  peaks exceeding
3650  m and 55 peaks  exceed  4260  m (Arno and
Hammerly 1984).  The estimated target population
size (N) in the Southern Rockies was  1609 (Table
5-5). The  median lake elevation (3264  m) in the
Southern Rockies was the highest of the five western
subregions surveyed. This subregion included the
lake with the highest elevation (ID number = 4E3-
053, 3883 m). Twenty percent of the lake population
exceeded 3552 m in elevation. Lakes in the Southern
Rockies were also the smallest, with a median area
of 3.5 ha and a Q4 area of 7.1  ha. They  were  also
the shallowest (Q-, = 2.8 m,  M =  5.1 m, and Q4 =
10.4 m) (Table 5-6).

The highest ANC values for any subregion in the
WLS-I were observed in the Southern Rockies. Only
4.6 percent of the lakes were estimated to have ANC
< 50 /jeq L"1 and only  39.4 percent of the lakes
had ANC < 200 //eq L"1 (Table 5-9). The median
ANC value (317.0 fjeq L"1) for  this subregion was
the highest in the West. Similarly, the Qi for ANC
(113.9 (Jeq  L"1)  was  the highest of any WLS-I
subregion (Table 5-16).

The Southern Rockies was one of two subregions
in the West (the other was the Northern Rockies,
4C) in which no  sampled lakes had pH <  6.0; the
minimum pH observed in this subregion was 6.02
(Volume II). The median pH was  7.60, the highest
estimated for any western subregion (Table 5-1 6).

Sulfate  concentrations  were also higher in  the
Southern Rockies than in any  other subregion (M
= 34.6 peq  L"1). This concentration was approxi-
mately five times greater than the median S04~2 in
California (6.6  //eq L"1), which  was the lowest
estimated for the West  (Table 5-16). Wind-blown
particles and dissociation  products of calcium
carbonate and gypsum from the Great Basin may
be the source of the Ca+2 and S04~2 in precipitation
samples collected in the Southern Rocky Mountains
(Reddy and Claassen 1 985; Creager 1 985). Although
the existing  NADP/NTN precipitation  monitoring
stations in the Southern Rocky Mountains are not
located  at high elevations, Hidy and Young (1986)
have calculated that for the  existing low elevation
sites, Ca+2 in precipitation is strongly associated with
SO4~2 (r = 0.92).  The  strength of this association
supports the assumption that some of the  lake water
S04"2 in the Southern Rocky Mountains is related
to atmospheric deposition of gypsum, especially in
alpine  areas. Lewis et al. (1984)  report  that there
is evidence that strong acid anion deposition occurs
in Colorado and  that strong acid precursors are
transported  greater  distances  and to  greater
elevation than are carbonates and cations such as
Ca+2  and Mg+2.  However,  Lewis  et  al. (1984)
evaluated data from 42 deposition collectors, only
one of  which was  above 3000  m. Clearly,  more
deposition data from higher elevations, more closely
related  to high elevation  lakes, is  required to
                     134

-------
understand this issue. Calcium concentrations in the
lakes of the Southern Rockies were the highest jn
the West with a median value of 233.1  /ueq L~1,
almost five  times the median Ca+2 value  of 43.2
//eq L"1 estimated for California (Table 5-16).

In the Southern Rockies, as elsewhere in the West,
there was a strong relationship  between ANC and
[Ca+2 + Mg+2] (Table 6-4). Median concentrations for
K+ (9.0 A
-------
Figure  7-5.    Locations of sampled lakes within geomorphic units in the Southern Rockies (4E), Western Lake Survey - Phase I.
              Lakes located outside the boundaries were not included in estimates for these subpopulations.
                                         WY
                                                       Medicine Bow
                                                       Range
                             Park
                             Range
                                                                     CO
                                                                       Front
                                                                       Range
                                                                                    4E
                                                                        S'a watch Uplift

                                                                          Sangre de Cnsto
                                                                          Uplift
               San Juan
               Mountains
                         136

-------
 Table 7-9.    Population Estimates and Statistics for Primary Variables by Geomorphic Unit in the Southern Rockies (4E),
            Western Lake Survey - Phase I
ANC < 50 /jeq L'1 ANC
Geomorphic Unit n*** N pc Nc Ncu QI
Park Range 21 134 0.058 8 16 82.0
Front Range 51 311 0.126 39 53 74.4
White River Uplift 10 172 0.000 0 - 130.7
Sawatch Uplift 11 252 0.000 0 - 103.2
San Juan Mountains 21 247 0.092 23 37 72.2
SO/2 > 50 /ueq L'1 SO4~
Geomorphic Unit n*** N pc Nc Ncu Q,
Park Range 21 134 0.222 30 78 13.3
Front Range 51 311 0.338 105 188 20.5
White River Uplift 10 172 0.174 30 78 8.2
Sawatch Uplift 11 252 0,391 99 181 35.6
San Juan Mountains 21 247 0.194 48 98 13.9
(pieq L~1) Ca+2 < 50/veq L 1 Ca+2 (jieq L 1)
M Q4 pc Nc Ncu Q, M Q4
345.11621.8 0.067 9 18 70.5 303.01486.1
337.5 564.3 0.105 33 45 63.9 222.6 478.1
509.7 1068.8 0.000 0 - 79.2 314.5 672.6
327.91216.9 0.000 0 - 100.3 306.6 977.9
124.3 233.3 0.037 9 18 67.5 129.3 369.8
2 (^eq L'1) DOC<2mgL"1 DOC (mg L'1)
M Q4 pc Nc Ncu Q, M Q4
24.6 264.9 0.290 39 56 1.44 3.40 5.59
35.2 90.2 0.647 201 271 0.84 1.51 3.37
10.3 22.4 0.426 73 142 0.87 2.12 3.90
42.6 118.6 0.645 163 268 0.66 1.37 2.29
21.9 44.8 0.740 183 279 0.68 1.09 1.88
n*** = number of lakes from which samples were obtained.
N = estimated number of l.akes.
gc = estimated proportion of lakes meeting the given criteria.
Nc = estimated number of lakes meeting the given criteria.
Ncu = 95 percent upper confidence limit for Nc.
Q! = first quintile (20th percentile).
M = median (50th percentile).
Q4 = fourth quintile (80th percentile).
( — ) = undefined.
Table 7-10.    Population Statistics for Selected Variables for Lakes in Geomorphic Units in the Southern Rockies (4E),
             Western Lake Survey - Phase I
Conductance
Mg + 2
(Meq L-1)
Na +
(jieq L^1)
K +
cr
SiO2
Fe
Geomorphic Unit
                            M
0^
M
Q,
                                                    M
Q,
                              M
                                    M
                                                                                        M
                                                                                                    M
Park Range
Front Range
White River Uplift
Sawatch Uplift
San Juan Mountains



Geomorphic Unit
Park Range
Front Range
White River Uplift
Sawatch Uplift
San Juan Mountains
13.0 38.7
11.9 37.5
14.2 45.7
15.8 42.2
11.5 20.2

Site Depth
(m)
Q, M
2.8 4.2
3.0 6.7
1.4 2.6
3.4 4.8
3.9 4.7
20.3 76.0
15.7 91.6
61.9 146.6
23.3 73.8
14.4 21.4

Lake Area
(ha)
Q, M
2.4 3.2
1.7 3.7
0.6 1.2
1.5 2.9
1.6 4.7
20.7
17.3
21.4
16.7
10.0


42.3
33.1
44.2
25.4
31.5


WA:LAa
QI
11.4
11.4
14.6
9.6
6.4
M
18.9
23.8
65.4
13.1
15.8
2.9 16.6
3.5 5.2
4.0 14.3
3.5 8.1
3.6 7.1
Secchi Disk
Transparency
(m)
Q! M
0.7 2.8
2.0 2.4
1.4 2.6
1.2 4.3
1.9 3.9
2.2
2.1
2.3
1.4
2.0









4.0
3.6
5.7
3.5
2.5









1.6
1.3
0.4
0.6
1.5









2.0
2.3
2.9
1.7
3.5









11.6
3.7
7.7
0.7
7.7









24.5
42.5
25.7
11.3
15.4









QI = first quintile (20th percentile).
M = median (50th percentile).
a Watershed area:lake area.

White River Uplift with ANC < 200 fjeq L"1 was 142
ha. Turk and Adams' (1983) study  included  lakes
less than 1 ha in area (mean = 0.4 ha) and these
smaller lakes are  included  in their estimate of 370
lakes  with alkalinity  < 200 /ueq L"1. Based on the
WLS-I, the estimated number of lakes greater than
or equal to 1 ha  with ANC < 200  fjeq L"1  in the
White River Uplift is 40. Excluding lakes smaller than
1 ha from the Western Lake Survey - Phase I target
                    population may affect population estimates if lakes
                    < 1 ha are prevalent.  In  the White River Uplift,
                    however, excluding small lakes had little effect on
                    the estimated  lake area  with  ANC < 200 yueq L"1.
                    The lakes in the White River Uplift were smaller and
                    shallower than in any other geomorphic unit.

                    In the Park Range, the bedrock consists of Cretaceous
                    sedimentary and Precambrian metamorphic mate-

                                             137

-------
rial. This area had the fourth highest median ANC
in the West. The Park Range also had the second
highest median DOC concentration (3.40 mg L"1)and
the  highest Q4 values for Ca+2 (1486.1 fjeq L"1) and
S04~2 (264.9 /jeq L"1) of any geomorphic unit in the
WLS-I (Table 7-9).

Lakes of the Sawatch Uplift  had the second highest
0.4 value for SO.T2 of  any geomorphic unit (118.6
fjeq  L"1). Acid neutralizing  capacity and  Ca+2
concentrations were relatively high in lakes in this
area. Of  the 11 lakes  sampled in this geomorphic
unit, none had ANC or  Ca+2 < 50 fjeq L"1 (Table 7-9).

7.6  Ionic Relationships for Lakes in the
Major Geomorphic Units

Ternary diagrams (Section 6.2.6) were developed for
the  major geomorphic  units  in each subregion.
Examples of these are given for the Sierra Nevada
(from 4A), the Oregon Cascades  (from  4B),  the
Bitterroots (from 4C),  the Wind River  Range  (from
4D), and the Front Range (from 4E; Figure 7-6). These
five geomorphic units contain  the  largest number
of dilute lakes  estimated for any geomorphic unit
in each  of the subregions. The  Sierra Nevada is
underlain by a granitic batholith. The Bitterroots are
on  the edge  of the Idaho  Batholith, whereas  the
Oregon Cascades are  extrusive andesite. The Wind
River Range is an uplift of Precambrian granites and
gneisses and the Front Range is composed  of
Precambrian granitic and metamorphic rocks.

Lakes in the Sierra Nevada, the Bitterroot Mountains,
and the Front Range are similar in composition with
respect to proportions of major ions (Figure 7-6), and
are largely dominated by Ca+2 and HC03~, although
lakes in  all three areas have moderate proportions
of Na+and K+. The proportion of Cl~ tends to be higher
in the Sierra  Nevada than in the Bitterroot Moun-
tains, and the proportion of Cl~ in the Front Range
is extremely  low. Lakes in  the Wind  River Range
have notably uniform ionic composition.

The Oregon Cascades also contain many lakes with
low ionic strength. These lakes have slightly higher
Mg+2, Na+, and Cl~ concentrations  and proportions
than  do lakes in  the Sierra  Nevada  and the
Bitterroots. The proportion and concentration  of
S04~2 in the  Oregon Cascades are especially low.
Other areas of the Cascades, such as the Wenatchee
Mountains, have lower relative proportions of Mg+2,
Na+, and Cl~ and higher relative proportions of Ca+2
and S04~2. Lakes in the Wind River Range and the
Front Range, in the Central and Southern Rockies,
have similar ionic composition (Figure 7-6).
                     138

-------
Figure 7-6.     Trilinear diagrams showing the relative abundance of major anions and cations for lakes in the largest geomorphic
               units in California (Sierra Nevada), the Pacific Northwest (Oregon Cascades), the Northern Rockies (Bitterroot
               Mountains), the Central Rockies (Wind River Range), and the Southern Rockies (Front Range), Western Lake
               Survey - Phase I. Ratios are expressed as percent of total ionic concentration, increasing from 0 to 100 percent
               along the axes in the direction of the arrows.
                                               .Ca                    Cl

                                                    Bitterroot Range
                    Ca                     Cl

                         Wind River Range
                                                                                          139

-------
                                          Section 8
          Comparison of Results from the Western and Eastern Lake Surveys
Phase I of the Eastern and Western Lake Surveys
(ELS-I and  WLS-I) collectively comprise the first
phase of the  National Lake  Survey, which was
designed to provide a data base that could be used
as a basis  for determining the present chemical
status and location of lakes in the United States that
are at  risk due to acidic deposition. The  results
discussed below relate to only  the areas covered by
sampling in each survey.

Data from  1592  lakes were used  to  generate
population estimates for eleven subregions in three
regions in the eastern United States - the Northeast,
Upper Midwest and Southeast.  Because the two
subregions  in the Southeast - the  Southern Blue
Ridge and Florida - contained lakes with substantially
differing characteristics, a  regional  data  summary
for the Southeast was not presented (Linthurst et
al. 1986; Overton et al. 1986; Kanciruk et al. 1986a).
The number of lakes used to  generate  population
estimates in the West was 719, distributed among
five subregions.  Based  on the number of lakes
sampled  and the estimated number  of lakes
comprising the map  population in  each region
(Linthurst et al. 1986; Section  2.3.2), the size of the
characterized  lake population was 7157 in  the
Northeast, 8575 in the Upper  Midwest, 286 in the
Southern Blue Ridge, 2138 in Florida, and 10,393
in the West. Because of the scales of maps available
on which map populations were identified, the target
population of lakes in the  ELS-I was restricted to
lakes > 4 ha, whereas the  target population of the
WLS-I included lakes > 1 ha in area. Approximately
44 percent of the  target population  of lakes in the
West were 1-4 ha, but there  was  little difference
in lake chemistry between these lakes and those
> 4 ha (Section 6.4.1). Consequently, the compar-
isons between the lakes  of  the  East  and West
discussed  here include  those lakes  in the West
between  1  and 4  ha. The summary that  follows
highlights the results of the two surveys with respect
to physical and chemical characteristics of regional
lake populations.

8.1   Physical Characteristics

Lakes in the western United States were much
higher in elevation than lakes  in the eastern United
States; the Qi value for elevation in the West (Table
8-1) was  higher than  all  Q4 values in the East
(Linthurst et al. 1986). Watershed areas for western
lakes were smaller than for lakes in the Northeast,
Upper Midwest, and Southern Blue Ridge, but were
slightly larger than for lakes in Florida; lake area
was smaller in the West than in all eastern regions.

Watershed area to lake  area  ratios were  much
greater in the West than in the Northeast, the Upper
Midwest, and Florida, but less than in the Southern
Blue Ridge. Despite their smaller size, western lakes
were deeper than eastern lakes. In the West, seepage
lakes were more common and reservoirs and closed
lakes were less common  than  in  the  Northeast.
Compared to  the Upper  Midwest and  Florida,
seepage lakes and closed  lakes were less common
and drainage lakes  and  reservoirs were  more
common in the West (Table 8-1).

Land  use/land  cover  differed  markedly between
watersheds in the East and West (unpublished data,
EPA Environmental Research Laboratory, Corvallis;
Section 5.3). Urban and agricultural land use was
much less prevalent in western watersheds.  No
western watersheds were  dominated by wetlands
whereas wetlands dominated in one to six percent
of the watersheds in six  eastern  subregions.
Rangeland was more common in the West than in
the East.  Barren land (exposed  rock), tundra, and
perennial snow or ice collectively dominated  12-58
percent of the watersheds in each western subre-
gion;  in contrast only one to four  percent of the
watersheds in three  eastern  subregions were
dominated by  barren  land and  no watersheds
contained tundra  or perennial  snow or ice. Total
forest cover was  similar  in  eastern and  western
watersheds; however, forests were  almost exclu-
sively evergreen  in the  West, whereas eastern
forests were  more likely  to  be mixed coniferous-
deciduous or deciduous.

8.2  Primary Variables

The  population distribution  for  acid neutralizing
capacity (ANC) for lakes in the western United States
was most  similar to that for lakes in the Northeast
(Figure  8-1).  Florida, however,  had a higher
percentage of lakes with  low  ANC and a higher
percentage of lakes with high ANC than did the West,
                                              140

-------
Table 8-1.    Physical Characteristics of Lakes Sampled During Phase I of the Eastern and Western Lake Surveys8
 Lake Elevation (m)

Q,      M     CL
                        Lake Area (ha)

                       ,       M      C
                                                                            Watershed Area (ha)   Watershed Area:Lake Area
                                                                                    M
                                                                                                            M
Northeast 1
Upper Midwest 2
Southern Blue Ridge 3A
Florida 3B
West 4
110
339
221
15
1746
307
408
265
21
2613
491
487
454
31
3237
8.0
7.6
5.2
6.5
2.2
16.7
14.8
10.8
17.3
4.6
62.7
46.1
64.3
44.1
11.7
90
56
182
35
45
271
177
682
115
118
1289
1023
5836
352
436
6.2
4.3
20.7
3.5
11.0
13.0
10.3
44.2
5.5
23.7
38.6
30.3
127.9
12.0
58.5
 Site Depth (m)

,     M      C
                                                           Hydrologic Lake Type (Percent of Population)
                                                        Drainage
                                        Seepage
                                                                                   Reservoir
                                                                 Closed
Northeast
Upper Midwest
Southern Blue Ridge
Florida
West
1
2
3A
3B
4
1.6
2.2
2.3
1.5
3.8
4.2
5.6
4.8
2.7
7.8
8.3
10.9
12.5
5.2
17.7
71
50
7
21
74
7
43
0
66
18
17
2
90
0
7
5
5
3
13
2
QI  = first quintile (20th percentile).
M  = median (50th  percentile).
Q4  = fourth quintile (80th percentile).
aData for Regions 1  and 2 and Subregions 3A and 3B are from Linthurst et al. (1986).
Figure 8-1.     Cumulative frequency distribu-
               tions [F(x>] for ANC (/jeq L~")
               and  pH  for  the  Northeast
               (	),  the Upper Midwest
               (	),  the Southern  Blue
               Ridge  f	).  and Florida
               (	). Eastern Lake Survey -
               Phase I (Linthurst et al. 1986),
               and the West (	). Western
               Lake  Survey - Phase I.
                  1.0
                 0.8-
                 0.6-
              FM
                 0.4-
                                               0.2-
                                               0.0
                                                          Northeast (1)
                                                          Upper Midwest (2)
                                                          Southern Blue Ridge (3A)
                                                          Florida (3B)
                                                          West (4)
                                                 -100
                                                                    200
                                                    I
                                                   400
                                                               I
                                                             500
                                                                                                          500
WOO
                                                                             ANC (ueq O
                                                       	Southern Blue Ridge (3A)
                                                       	Florida (3B)
                                                            West(4)                     -'I I
                                                                                    r     I
                                                                                           141

-------
as reflected by the Qi and median ANC values (Figure
8-2).

With the exception of one lake associated with hot
springs, there were no acidic lakes (ANC < 0 //eq L~1)
sampled in the WLS-I (Figure 8-3). In contrast, 4.6
percent of the lakes in the Northeast, including 10.7
percent in the Adirondacks, were acidic. Of the lakes
in the Upper  Midwest,  1.7 percent had ANC < 0
yueq L~1, ranging from none sampled in Northeastern
Minnesota to 9.8  percent in  the  Upper Peninsula
of Michigan (Linthurst et al. 1 986). None of the lakes
in the Southern Blue Ridge was estimated to have
ANC  < 0 yueq  L"1, but  22 percent of the lakes  in
Florida were estimated to be acidic.

The percentage of lakes  estimated to have ANC < 50
/ueq L~1 in  the West was similar to that estimated
for the Northeast, but was roughly one-half that for
Florida (Figure 8-3). The percentage of lakes in the
West with ANC <  200 /ueq  L"1 (66.6%) was the
highest estimated for any region. Estimates of lakes
in this ANC class were 60 percent for the Northeast,
55.1 percent for Florida, 41.4 percent for the Upper
Midwest and  34.3 percent for the Southern Blue
Ridge. The percentages of lakes with ANC  < 200
                                          fjeq L"1 in California, the Pacific Northwest and the
                                          Central Rockies were greater than for any subregion
                                          sampled  in the eastern  United States. In the East,
                                          lakes with  low ANC were generally much smaller
                                          than those with high ANC (Linthurst et al. 1986).
                                          In contrast, lakes in California were the most dilute
                                          and had the lowest values of ANC. California lakes
                                          were generally larger and were among the deepest
                                          sampled in the West.
                                          Most lakes surveyed in the western and eastern
                                          United States had pH > 5.0 (Section 5.4.2; Linthurst
                                          et al. 1986). Values for pH were generally  higher
                                          in the West than in other areas, but, as was observed
                                          for ANC, considerable  variation in the  population
                                          distributions existed (Figure 8-1). The most marked
                                          differences were observed at low pH values. First
                                          quintile values for the West and the Southern Blue
                                          Ridge were similar and were higher than in any other
                                          area (Figure 8-2). Approximately one  percent of the
                                          lakes in  the West and 0.4 percent in the Southern
                                          Blue Ridge had pH < 6.0 (Figure 8-3). In contrast,
                                          12.9  percent of the lakes in  the Northeast, 9.6
                                          percent of the lakes in the Upper Midwest and 32.7
                                          percent of the lakes in Florida  had pH < 6.0.
Figure 8-2.
  Concentrations of ANC (neq L'''), pH. and concentrations of sulfate (/jeq L 1) for lakes in the Northeast (Region
  1). Upper Midwest (Region 2). Southern Blue Ridge (Subregion 3A), and Florida (Subregion 3B), Eastern Lake
  Survey - Phase I (Linthurst, et al. 1986). and West (Region 4).  Western Lake Survey - Phase I. First quintile
  fQi. 20th percentile), median (M, 50th percentile). and fourth quintile (Qt, 80th percentile) values are shown
       o-
       to
7200



7000



 800



 600



 400



 200



   0

-100
               _ O
                          O
                          -H
H—h
 2 3A 3B 4

 142
                                      ej.
                                         8.0
                                         7.5
                                         7.0
                                         6.5
                                         6.0
                                         5.5
                     5.0
                           1
  04

  M
                                                           O  _
H — I — I — I — H
 1  2 3A 3B 4

                                                                        280
                                                                        240
                                                                        200
                                                                        160
                                                                        120
                                                                         80
                                                                         40
                                                                                  O  -•
H—I—I—h
 1  2  3A 3B 4

-------
Figure 8-3.   Population estimates for the
            percent of lakes with concen-
            trations ofANC <0,< 50. and
            < 200 ueq L~\ and pH < 6.0,
            for the Northeast, Upper Mid-
            west, Southern Blue  Ridge,
            and Florida, Eastern Lake Sur-
            vey - Phase I (Linthurst et al.
            1986). and West, Western Lake
            Survey - Phase I.
                                       T
ANC<0
     ueq L'\
ANC < 50
                                  ANC < 200 fjeq L
                                         pH < 6.0
                                                          20
                            B Northeast (1)
                            • Upper Midwest (2)
                            D Southern Blue Ridge (3A>
                            Q Florida (3B)
                              West (4)
                                                 T
                                                                    40       60
                                                                        Percent
                                                 80
                                                         100
A  striking contrast  among the  distributions  for
primary variables between lakes in the western and
eastern United States was observed f or S04~2 (Figure
8-4). The distributions for S04~2 in the West closely
paralleled that observed for the Southern Blue Ridge
and, in both these areas, S04~2 concentrations were
much lower than in the Northeast, Upper Midwest
and Florida. The Q4 value for S04~2 in  lakes in the
West was 38.7 yueq L~1, less than one-half the Qi
value in the Northeast (Figure 8-2). In the Northeast,
97.3 percent of the lakes  had  S04"2 > 50 ,ueq L~1,
while only 13.5 percent of  the lakes in the West
had S04"2 > 50 /ueq L"1 (Figure 8-5).

The population distributions for Ca+2 in western and
eastern lakes generally paralleled those  for ANC
(Figure  8-4).  At  lower Ca+2  concentrations, the
distribution in the West was  most similar to that
in  the Southern Blue  Ridge, but, at higher concen-
trations, the distribution in the  West was practically
indistinguishable from that for the Northeast. This
comparison is also reflected in the median and Qi
values for the West  and the Southern Blue Ridge
and in the Q4 value for the West and the Northeast
(Figure 8-6). More lakes in the West had Ca+2 < 50
/ueq L"1 than in any area in the eastern United States
(Figure 8-5).

Extractable Al in clearwater lakes (true color < 30
PCU)  was  low among most areas surveyed.  The
distribution for extractable Al for western lakes was
intermediate  between that for the Southern Blue
Ridge and those for the Northeast, Upper Midwest
             and Florida; the distributions in the latter three areas
             were indistinguishable (Figure 8-7). The lowest Qi
             value was observed for clearwater lakes in the Upper
             Midwest, while the highest occurred in the Northeast
             (Figure 8-6);  the median  value in  the  West  was
             similar to that in the Upper Midwest. Less than one
             percent of the clearwater  lakes in the West  had
             extractable Al > 50 /JQ L~1 and no lakes in  this
             category were sampled in the Southern Blue Ridge
             (Figure 8-5).

             Dissolved organic carbon differed markedly among
             eastern and western lakes. The distribution for DOC
             in  the  West  most closely paralleled that in the
             Southern Blue Ridge (Figure 8-7). Concentrations of
             DOC in these  areas were markedly less than in other
             areas surveyed (Figure 8-6).  Nearly 70  percent  of
             the lakes in the West had DOC < 2 mg L~1, whereas
             less than ten percent of the lakes in the Northeast
             and Upper Midwest had DOC < 2  mg L"1 (Figure
             8-5). The West also contained very few systems with
             high DOC (> 6 mg L~1), similar to the Southern Blue
             Ridge,  but unlike the Upper  Midwest and Florida
             where  more than 60 percent of the  lakes had DOC
             > 6 mg L"1 (Figure 8-5). The West contained many
             more clearwater lakes with low  DOC than did areas
             in  the eastern  United States, and  Secchi  disk
             transparencies were correspondingly  higher (Sec-
             tion 5.6.2; Linthurst et al. 1986).

             Many lakes in the West had low ionic concentration
             (Section 5.9). The lowest Qi value (20th  percentile)
             for conductance  in the East was  20.7 fjS cm"1
             (Linthurst et  al.  1986).  In contrast, 20  percent  of
                                                                           143

-------
Figure 8-4.    Inverse cumulative frequency
              distributions [1 -F(x)] for sulfate
              f/jeq L~^t  and cumulative fre-
              quency distributions [F(x)J for
              calcium (/jeq L~^) for the North-
              east (	),  Upper  Midwest
              (	), Southern Blue Ridge
              I	), and  Florida  f	). 1-nxc>
              Eastern Lake Survey - Phase I
              (Linthurst et  al.  1986), and
              West  (	). Western Lake
              Survey - Phase I.                    0.2 -
                                        Northeast (1)
                                        Upper Midwest (2)
                                        Southern Blue Ridge <3A)
                                        Florida (3B)
                                        West 14)
                                                 1.0-
                                             F(x)
0.8-


0.6-


0.4-


0.2-
                                                 0.0-
                                                                             720          180

                                                                               Sulfate ffjeq L'^)
                                                      240
                                                                  300

                                                                                        Northeast (1)
                                                                                        Upper Midwest (2)
                                                                                        Southern Blue Ridge (3A)
                                                                                        Florida (3B)
                                                                                        West (4)
               200         400          600

                              Calcium ffjeq L~^)
                                                                                                      800
                                                                                                                   WOO
Figure 8-5.     Population  estimates for  the
               percent of lakes  with concen-
               trations  of  calcium  <  50
               Lieq L~\ sulfate > 50 iieq  L~\
               DOC <2 mg L~\ DOC > 6 mg
               L~\ and extractable aluminum
               > 50 fjg i~1 for the Northeast.
               Upper Midwest, Southern Blue
               Ridge,  and Florida. Eastern
               Lake Survey - Phase I (Linthurst
               et al. 1986) and West, Western
               Lake Survey - Phase I.
                                   I     '     T     '     I
                                53 Northeast (1)
                                • Upper Midwest (2)
                                0 Southern Blue Ridge (3A)
                                  Florida (3B)
                                D  West (4)
                                            DOC > 6 mg L
                                          Ext.A/>50/jgL~-
                                                                                                                     100
                          144

-------
Figure 8-6.    Concentrations of calcium (/MO, L 1A extractable aluminum ffjg L'^) in clearwater lakes, and DOC (mg L'^) for
             lakes in the Northeast (Region 1), Upper Midwest (Region 2), Southern Blue Ridge (Subregion 3A), and Florida
             (Subregion 3B). Eastern Lake Survey - Phase I (Linthurst et at. 1986), and West (Region 4). Western Lake Survey
             - Phase I. First quintile (Q,. 20th percentile), median (M. 50th percentile), and fourth quintile (Q*. 80th percentile)
             values are shown (O—I—D ).
                         Q,  M  Q4
1000
       800  -
       600  -
    a
       400  -
       200  -
C
-
~ 0

'_
<


•> <
p



C
' <
	 1 — I — I

[


p
> ,
p
-
C


0 C
p
'_
. -
>
I 	
                                     I
                               §
                              I
               1  2 3A 3B 4
20
18
16
14
12
10
g

6
4
2

0

r
-
-
~ C
-


—
-
—
(


p

C




}

p



J
1 J
p
-
-
—
-

D
—
-
_
} .
k 7 , O

1 2 3A 3B 4
                                                                       14
                                                                        12
                                                                       10
                                                                     ~J
                                                                     I
                                                                     O
                                                                     O
                                                                              • • Q
                                                                       1  2  3A 3B 4
the lakes in the West had conductance^ 8.3//S cm"1,
and the  same  percentage in  California  had
conductance < 4.9 fjS cm"1 (Section  5.6.5). These
dilute western lakes had very low concentrations of
dissolved constituents, such as calcium, magnesium,
sulfate, and nutrients  (Section 5.9).
8.3  Associations Among Variables
Many lakes in the West occurred at relatively high
elevations, particularly in  the  Sierra Nevada  in
California, the Central Rockies,  and the  Southern
Rockies (Section 5.3). There was a moderately strong
inverse relationship between elevation and ANC in
the Sierra Nevada and Central Rockies,  but there
was little or  no  relationship in the other areas
(Section 6.4.2). The only area in the East where a
strong positive relationship between elevation and
ANC was found was the Adirondacks, although few
eastern areas had appreciable elevational  gradients.
                                             Seepage lakes had the lowest values of Ca+2 and
                                             ANC in the Upper Midwest and in the  Northeast;
                                             drainage lakes had the lowest ANC in the Northeast
                                             (Linthurst et al.  1986). In  California and the Pacific
                                             Northwest, drainage lakes had higher ANC and Ca+2
                                             than did seepage lakes, but the reverse was true
                                             for the subregions in the Rockies (Section 6.4.3).

                                             The  sums of Ca+2 and Mg+2 for lakes in the western
                                             subregions were much lower than for lakes in the
                                             eastern subregions (Section 6.2.2; Linthurst et al.
                                             1986).  Lakes  in several  areas of the  West  had
                                             concentrations of base cations that were comparable
                                             to some of the most  dilute  lakes  in  the world
                                             (Armstrong and Schindler  1971).

                                             Most lakes in  the West were dominated by calcium
                                             bicarbonate (Section 6.2.5). In some lakes, Na+ was
                                             the  dominant cation and HC03" or  CI" was the
                                             dominant anion. A few lakes were sampled that had
                                             Ca+2 and S04~2  as the dominant ions. In contrast.
                                                                           145

-------
Figure 8-7.   Inverse cumulative frequency
            distributions  [1-F(x)]  for
            extractable aluminum fog L~^)
            in clearwater lakes, and cumu-
            lative  frequency distributions
            [Ffxj] for DOC (mg i"V for the
            Northeast  (	).  Upper
1.0
0.8-
0.6-
            Midwest (	). Southern l-F(x)
            Blue Flidge (	), and Florida
            (	), Eastern Lake Survey -
            Phase I (Linthurst et al. 1986)
            and  West f	). Western
            Lake Survey - Phase I.
0.4-
                                        0.0
	 Northeast (1)
- — Upper Midwest (2J
	Southern Blue Ridge (3Aj
	Florida (3B)
     West (4)
                                                     10
                                                            Extractable Aluminum (ug L~
                                                                      	Southern Blue Ridge (3A)
                                                                      	Florida (3B)
                                                                           West(4)
                                                                  DOCfmgL"1)
S04"2 was the dominant anion in 20 percent of the
lakes in three of five subregions in the Northeast,
in the  Upper Peninsula  of  Michigan,  and  in
Northcentral Wisconsin  (Linthurst  et  al. 1986).
Sulfate  was also high in  Florida lakes.Many more
lakes in the East had Na+ and Cl~ as major ions.

In general, lakes in the West had lower  concentra-
tions of Ca+2, Mg+2, S04"2, and  DOC than did lakes
in the East.  Many very dilute lakes were found,
particularly in the Sierra Nevada, the Cascades and
the Wind  River Range. These lakes have very low
concentrations  of  dissolved constituents  and are
highly transparent.

8.4  Discrete Subpopulations of Lakes
in the East and West
By combining lake populations from entire regions,
important differences among groups containing the
most low ANC lakes are sometimes obscured. Four
subpopulations containing the highest proportion of
acidic  lakes were  selected from  the East for
comparison with five previously described geomor-
phic units in the  West  (Sections  7.1-7.5).  The
           Adirondack Park  represents  the estimated  lake
           population within  the park boundary. The eastern
           half of the Upper Peninsula of Michigan describes
           those lakes in the  Peninsula east of longitude 87°.
           Northcentral Wisconsin is Subregion 2C as shown
           in Linthurst et al. (1 986). Northcentral Florida is also
           described  in  Linthurst  et  al.  (1986)  and closely
           coincides with the  area known as the Trail Ridge.

           The population statistics for pH clearly demonstrate
           the  major differences  between the eastern and
           western subpopulations, with Qi values differing by
           more than one pH  unit (Figure 8-8). However, when
           compared  on  the  basis  of ANC,  the eastern and
           western subpopulations are similar (Figure 8-8). The
           large interquintile differences  in ANC for lakes in
           Michigan,  Wisconsin, and Florida compares closely
           with those in the Bitterroot Mountains and the Front
           Range. Small interquintile differences for ANC were
           observed in the Adirondack Park, the Sierra Nevada,
           Oregon Cascades,  and Wind River Range. Such low
           variability suggests  a  relatively high  degree of
           chemical and  physical homogeneity within  these
           geomorphic units. Differences in S04 2 concentra-
           tions in the eastern and western subpopulations
                     146

-------
Figure 8-8.    Concentrations of ANC f/jeg L'^) and pH in selected subpopulations of lakes located in major geomorphic units
             in the Northeast, Upper Midwest, and Florida. Eastern Lake Survey - Phase I (Linthurst et al. 1986) and West,
             Western Lake Survey - Phase I. First quintile (Q,. 20th percentile), median (M, 50th percentile). and fourth
             quintile (Qt. 80th percentile) values are shown (O—t—O)
                                                   Qi  M Q4
    Adirondack Park,
      New York

    Eastern Upper Penin-
      sula of Michigan

    Northcentral
      Wisconsin

    Northcentral
      Florida

    Sierra Nevada,
      California

    Oregon Cascades
    Bitter root Mtns.,
     Idaho/Montana

    Wind River Range,
     Wyoming

    Front Range,
     Colorado
„*•« jv


62 54U
loo J4BU
50 963 ~
114 2119-
42 443 ~
J/ 283
47 884 -

(
i | i | i | i | i | i




OH 	 D
OH 	 D
. 1 , 1 1 1 1 1 I 1 1 1
3 200 400 600 SOO 1000 1200 I
I I I




o — ho
o-H-a
I I I
5 6 7 8
                                                     ANC(ueqL")
PH
were also striking  (Figure  8-9). Only in the Front
Range  in Colorado did  concentrations approach
those observed in the East.

Large  differences  in DOC concentrations between
lakes in the East and West were also observed (Figure
8-9). Dissolved organic  carbon can be used to
estimate  organic acid concentration, which may be
much  more  important in the chemistry of eastern
lakes than in western lakes. As expected, the pattern
observed  for Ca+2  concentrations in these subpop-
ulations (Figure 8-9) closely mirrors that  for ANC.
                                                                                 147

-------
Figure  8-9.    Concentrations of sulfate (fieq L'1). calcium (/jeq L  '), and DOC (mg £~V '" selected subpopulations of lakes
              located in major geomorphic units in the Northeast,  Upper Midwest, and Florida. Eastern Lake Survey - Phase I
              (Linthurst et al. 1986) and West, Western Lake Survey -  Phase I.  First quintile (Q,, 20th percentile).  median
              (M, 50th percentile), and fourth quintile (Qt, 80th percentile) values are shown ( Q	1—D )
                                                                                       Q,  M  Q<
                      /i"*  N

  Adirondack Park.    127   1091-
    New York

   Eastern Upper Penin- 62   540 .
    sula of Michigan
   Northcentral
     Wisconsin

   Northcentral
     Florida

   Sierra Nevada,
     California
153  1480-


 50   963 -


114  2779-lK)
   Oregon Cascades    42   443-D
   Bitterroot Mtns ,
     Idaho/Montana

   Wind River flange,
     Wyoming

   Front Range,
     Colorado
 37   283-
 47   884-
 51   311-
                 o-f-Q
0-H3
                419
                                       I    i   I
1 1 1 1 1 1 1
1020

876
H-Q
H-a
0+-D
1 1 I 1 1 1 I 1
' 1 ' 1 ' 1 *


H-o
o-K3
o4-n
o-| 	 a
i 1 i 1 i 1 i
                                0    100    200    300

                                   Sulfate (yeqL"11
                                    0    200    400   600   800  0
                                          Calcium (fieq i"V
                                                           4      8     12
                                                           DOC(mgL^)
                                                                               16
                         148

-------
                                           Section 9
                                          Conclusions
Within the areas  of  the  western United  States
sampled and covered by the design of the Western
Lake Survey it is concluded that:

• More than  one-fourth  (26.6 percent) of the
  western  lakes  are dilute  (conductance <  10
  fjS cm"1). The areas sampled in the western United
  States contain a much higher percentage of lakes
  with low concentrations of  dissolved substances
  than do areas sampled in three regions (Northeast,
  Upper Midwest,  and Southeast) in the  eastern
  United States. The lowest Qi value (20th percen-
  tile) for  conductance measured in the East was
  20.7 fjS cm"1. In contrast, 20 percent of the lakes
  in the West had conductance values < 8.3 fjS cm"1
  and the minimum measured conductance in the
  West was 1.6 /uS cm"1.

• Calcium concentrations in the West varied widely
  among subregions;  56.5 percent of the lakes  in
  California had calcium < 50 /jeq L"1 whereas only
  3.4 percent of the lakes in the  Southern Rockies
  were below  this  criterion.  Overall, the  median
  concentration of calcium for lakes  in  the  West
  was low (92.4 fjeq L"1) compared to the Northeast
  (177.4 fjeq L"1), the Upper Midwest (238.2 fjeq L"1),
  and Florida (238.3 fjeq L"1). Lakes in the Southern
  Blue Ridge had a comparable  median value for
  calcium (104.7 //eq L"1).

• Median  sulfate concentrations were  extremely
  low throughout the West, ranging from 6.6 yueq L"1
  in California  to  34.6 fjeq  L"1  in the  Southern
  Rockies.  In comparison, median concentrations
  for  lakes in the East were 1 15.4 fjeq L"1 for the
  Northeast, 57.1 fjeq L"1  for the Upper Midwest,
  31.8 fjeq L"1 for the Southern Blue Ridge, and
  93.7 fjeq L"1  for Florida.

• Extractable aluminum concentrations in clear-
  water lakes were very low throughout  the West.
  Only 0.2 percent of the lakes in the West had
  extractable aluminum concentrations exceeding
  50  //g L"1 compared to 5.5 percent of the lakes
  in the Northeast, and 7.4 percent of the  lakes  in
  Florida. These results are, again, similar to those
  for  the  Southern  Blue  Ridge where  no  lakes
  exceeded this criterion.
• Concentrations of dissolved organic carbon (DOC)
  were also  low throughout the West; only 5.4
  percent of the lakes in the West had DOC values
  exceeding 6 mg  L"1, whereas this criterion was
  exceeded  for 26.4  percent of the lakes in the
  Northeast,  62.9  percent in the Upper Midwest,
  and 68.9 percent in Florida. Again, the Southern
  Blue Ridge shows a similarity to  western  lakes
  with 6.1 percent of the lakes with DOC > 6 mg L"1.

• pH values were  not low in the West, where 99
  percent of lakes had values greater than 6.0.
  Median pH values for the subregions ranged from
  6.94 in California to 7.60 in the Southern Rockies.

• Of the subregions sampled in the West, California
  had  the largest  number (881) and  percentage
  (36.7%) of lakes with low ANC (<  50 fjeq L"1);
  followed by the Pacific Northwest  (333 lakes;
  19.5%). The Southern Rockies had the smallest
  number and the  lowest percentage of lakes with
  low ANC (74 lakes; 4.6%).

• The  lakes in  wilderness areas had  much  lower
  concentrations of ANC with a median value  of
  91.4 fjeq L"1 compared to 282.7 fjeq L"1 for non-
  wild.erness lakes.

• The Sierra Nevada (the area which contained the
  majority of lakes sampled in California) had the
  largest estimated number of low ANC lakes (834)
  and  dilute  lakes (1311)  of any geomorphic unit
  in the West.

• No lakes sampled in the West had ANC <0//eq L"1,
  with the exception of one  lake associated with
  a hot spring.

• Evaluation of ion ratios,  the relationship  of base
  cations to ANC,  and the positive relationship  of
  calcium to sulfate  suggest that chronic regional
  lake  acidification was not evident in the moun-
  tainous areas of the West.
                                              149

-------
                                       Section 10
                                       References
 1.  Aimer, B., W. Dickson, C. Ekstrom, and E.
    Hornstrom. 1978. Sulphur Pollution and the
    Aquatic Ecosystem, pp. 271-311. In: Sulphur
    in the Environment:  Part  II, Ecological
    Impacts,  J.O.  Nriagu, ed. John Wiley and
    Sons, New York, New York.

 2.  Alt,  D. D. and D. W.  Hyndman.  1972.
    Roadside Geology of the  Northern Rockies.
    Mountain Press, Missoula, Montana. 280 pp.

 3.  Alt, D. D. and D. W. Hyndman. 1978. Rocks,
    Ice and Water.  The Geology of  Waterton
    Glacier Park.  Mountain  Press,  Missoula,
    Montana. 104 pp.

 4.  Altshuller, A.  P. and  R.  A. Linthurst, eds.
    1984.  The Acidic Deposition Phenomenon
    and Its Effects. Critical Assessment Review
    Papers, Volume II: Effects Sciences.  EPA-
    600/8-83-016 BF, U.S. Environmental
    Protection Agency, Washington, D.C.

 5.  Anderson, J. R., E. E. Hardy, J.  T. Roach,
    and R. E. Whitman. 1976. A Land Use and
    Land  Cover  Classification System for Use
    with Remote Sensor Data. U.S. Geological
    Survey Professional Paper 964. 28 pp.

 6.  Armstrong, F. A. J.  and D.  W. Schindler.
    1971.  Preliminary Chemical Characteriza-
    tion  of Waters in  the Experimental   Lakes
    Area, Northwestern Ontario. J. Fish. Res. Bd.
    Canada 28:171-187.

 7.  Arno,  S.  F.  and R.  P.  Hammerly.   1984.
    Timberline.  Mountain and  Arctic Forest
    Frontiers.  The Mountaineers,  Seattle,
    Washington. 304 pp.

 8.  Arno, S. F. 1969. Glaciers in the  American
    West. Natural History 78:84-89.

 9.  Baker, J. P.  and  C. L  Schofield.   1982.
    Aluminum Toxicity to Fish in  Acidic Waters.
    Water, Air, and Soil Poll. 18:289-309.

10.  Baldwin, E. M. 1981. Geology of Oregon, 3rd
    Edition. Kendall/Hunt Publishing  Company,
    Dubuque, Iowa. 170 pp.
11.   Baron, J.  1983. Comparative Water Chem-
     istry  of  Four Lakes  in  Rocky  Mountain
     National  Park. Water  Resources  Bulletin
     19:897-902.

12.   Bonoff, M. B. and A.  W.  Groeger.  1987.
     National  Surface  Water Survey, Western
     Lake  Survey  - Phase  I, Field Operations
     Report. Internal  Report. EPA/600/8-87/
     018.  Environmental Monitoring Systems
     Laboratory, U.S.  Environmental  Protection
     Agency, Las Vegas, Nevada, 26 pp.

13.   Bortelson, G.  C., R. T. Wilson,  B. L. Fox-
     worthy. 1977. Water quality effects on Baker
     Lake of recent volcanic activity at Mt. Baker,
     WA.  U.S. Geological Survey  Professional
     Paper 1022-B. 30pp.

14.   Bradford,  G. R., F. L. Bair, and V. Hunsacker.
     1968. Trace and Major Element Content of
     170 High  Sierra Lakes in California. Limnol.
     Oceanogr. 13:526-530.

15.   Brakke, D. F. and T. J. Loranger. 1986. Acid
     Neutralizing Capacity of Lakes  in the North
     Cascades Area of Washington State. Water,
     Air, and Soil Poll.  30:1045-1053.

16.   Brakke, D. F. and D. A. Waddell. 1985. Water
     Chemistry  in the North  Cascades  of
     Washington, pp. 79-101. In: Acid Precipita-
     tion in the Pacific Northwest, J. H. Baldwin,
     ed.. University of Oregon, Eugene, Oregon.

17.   Brezinski, D.  P.  1983.  Kinetic, Static, and
     Stirring Errors of  Liquid Junction Reference
     Electrodes. The Analyst 108:425-442.

18.   Busby, M. W. 1966. Annual Runoff in the
     Conterminous United States. Atlas HA-212
     Hydrologic Investigations. U.S. Department
     of the Interior. U.S. Geological Survey.

19.   Chen, C.  W.,  S. A. Gherini, N.  E. Peters, P.
     S.  Murdock,  R.  M.  Newton, and  R.  A.
     Goldstein. 1984. Hydrologic  Analysis of
     Acidic and Alkaline Lakes. Water Resources
     Research 20:1875-1882.
                                            750

-------
20.  Chronic, H. 1986. Pages of Stone. Geology
     of Western National Parks and Monuments.
     2:  Sierra  Nevada, Cascades, and Pacific
     Coast. The Mountaineers, Seattle, Washing-
     ton. 182pp.

21.  Creager, C. S. 1985. Characterization of the
     Influence of Soil Particulateson Precipitation
     Chemistry at Five Sites in the Western United
     States.  Report No.  0785-224-F,  Kilkelly
     Environmental Associates, Raleigh, North
     Carolina. 111  pp.

22.  Drever,  J. I.  1982.  The  Geochemistry of
     Natural  Waters.  Prentice Hall, Englewood
     Cliffs, New Jersey. 388 pp.

23.  Drever,  J. I.  and D.  R.  Hurcomb.  1986.
     Neutralization  of  Atmospheric Acidity  by
     Chemical Weathering in an Alpine Drainage
     Basin in the  North  Cascade  Mountains.
     Geology 14:221-224.

24.  Driscoll, C. T., J. P. Baker, J. J. Bisogni, and
     C.  L. Schofield. 1980.  Effect of Aluminum
     Speciation on Fish in Dilute Acidified Waters.
     Nature 284:161-164.

25.  Drouse, S. K., D. C. Hillman, L. W. Creelman,
     and  S. J.  Simon. 1986.  National  Surface
     Water Survey, Eastern Lake Survey - Phase
     I, Quality Assurance Plan. EPA/600/4-86/
     008.  Environmental  Monitoring  Systems
     Laboratory. U.S.  Environmental Protection
     Agency, Las Vegas, Nevada. 211 pp.


26.  Duncan, L. C. and W. Ausserer. 1984. Acid
     Precipitation in the Washington  Cascades,
     Phase II.  Final  Report  Contract  A-119-
     WASH, Washington Water Research Center.
     Pullman, Washington.

27.  Duncan, L. C. 1985. The Status (with Respect
     to Acidification or Potential Acidification) of
     Twenty-six Lakes Lying  in the  Cascades
     South of Chinook Pass. Final  report to the
     Washington State Department of Ecology.
28.  Eilers,  J.  M.,  D.  J.  Blick,  Jr.,  and M.  S.
     DeHaan. 1987a. Validation  of the Western
     Lake Survey -  Phase I Data Base. Corvallis
     Environmental  Research Laboratory,  U.S.
     Environmental Protection Agency, Corvallis,
     Oregon.

29.  Eilers, J. M., P. Kanciruk, R. A. McCord, W.
     S. Overton, L. Hook, D. J. Blick, D. F. Brakke,
     P. E. Kellar, M. S. DeHaan, M. E.  Silverstein,
     and D. H. Landers.  1987b. Characteristics
     of  Lakes in the  Western  United  States.
     Volume  II: Data Compendium for Selected
     Physical and Chemical Variables. EPA/600/
     3-86/054b, U.S. Environmental Protection
     Agency, Washington, D.C.

30.  Environmental  Data Service. 1983. Clima-
     tography of United  States #81  (by State),
     Monthly Normals of Temperature, Precipita-
     tion, and Heating and Cooling Degree Days,
     1951-1980. National Oceanic and Atmos-
     pheric Administration, U.S. Department of
     Commerce.

31.  Fountain, J., D. T. Hoff, and C. C. MacLeod.
     1986. Aquarius Programmers  and Users
     Guide.  Volumes  I  and II. Environmental
     Monitoring Systems Laboratory, U.S. Envir-
     onmental Protection Agency,  Las Vegas,
     Nevada. 323 pp.

32.  Fox, C. A.,  G.  R. Bradford, A. L. Page, and
     A.  A.  Elseewi. 1982. Evaluation of Acidity
     and Chemical Composition of High Elevation
     Lakes in California. Presented at Acid Rain:
     a North  American Challenge - Soil Conser-
     vation Society of  America  Symposium.
     October 25-28, 1985, Burlington, Vermont.

33.  Franklin, J.  F.  and C. T. Dyrness.  1973.
     Natural Vegetation of Oregon and Washing-
     ton. USDA  Forest Service Technical Report
     PNW-8.  417pp.

34.  Galbraith, A. F. 1984. The Acid Deposition
     Potential for the Streams and Lakes of the
     Wind  River Mountains,  pp. 275-272. In:
     Workshop Proceedings - Air Quality and Acid
     Deposition  in  the  Bridger  and Fitzpatrick
     Wildernesses.  USDA Forest Service, Inter-
     mountain Region, Ogden,  Utah. 292 pp.


35.  Galbraith, A. F. 1 985. Acid Deposition in the
     Wind  River Mountains. In: Proceedings of
     Watershed Management  in the Eighties -
     New Directions, Amer. Soc. Civil Engrs. April
     30-May 1, 1 985, Denver, Colorado. 18 pp.

36.  Galloway, J.  N.,  G. E. Likens, and  E. S.
     Edgerton. 1976. Acid Precipitation  in the
     Northeastern United States: pH and Acidity.
     Science 194:722-724.

37.  Galloway, J. N. and E. B. Cowling. 1978. The
     Effects  of  Precipitation  on Aquatic  and
     Terrestrial Ecosystems-a  Proposed  Precipi-
     tation Chemistry Network. J. Air Poll. Cont.
     Assoc. 28:229-235.
                                                                       151

-------
38.  Galloway, J. N. and P. J. Dillon. 1983. Effects
     of Acidic  Deposition: the Role of Nitrogen,
     pp. 145-160. In: Ecological Effects of Acidic
     Deposition.  Report PM  1636. National
     Swedish Environment Protection Board.

39.  Garrets, R. M.  and F. T. Mackenzie. 1967.
     Origin  of the  Chemical Compositions  of
     Some Springs  and Lakes.  In: Equilibrium
     Concepts in Natural Water  Systems, R. M.
     Garrels, ed. Amer. Chem. Soc. Adv. Chem.
     Ser. 67:222-242.


40.  Gibson, J. H., J. N. Galloway, C. Schofield,
     W. McFee, R. Johnson, S. McCarley, N. Dise,
     and  D. Herzog.  1983.  Rocky  Mountain
     Acidification  Study.  U.S.  Department  of
     Interior, Fish  and  Wildlife Service, FWS/
     OBS-80/40.17, Air Pollution and Acid Rain
     Report No. 17.  137 pp.

41.  Gilliom, R.  J.  and G.  C. Bortelson. 1983.
     Relationships between Water Quality and
     Phosphorus Concentrations for Lakes of the
     Puget  Sound  Region, Washington.  U.S.
     Geological Survey Open File Report 83-255.
     26pp.

42.  Gjessing, E. T.  1976. Physical and Chemical
     Characteristics of Aquatic Humus. Ann Arbor
     Science Publishers, Ann Arbor, Michigan.
     120 pp.

43.  Gorham,  E., S. J. Eisenreich, J. Ford, and
     M. V. Santelmann. 1985. The Chemistry of
     Bog  Waters,  pp.  339-363.  In: Chemical
     Processes in Lakes, W. Stumm., ed., John
     Wiley and Sons, New York, New York.

44.  Henriksen, A. 1980. Acidification of Fresh-
     waters—a Large-Scale Titration, pp. 68-74.
     In: Ecological  Impacts of Acid Precipitation,
     D. Drablos and A. Tollan, eds. SNSF Project,
     Oslo, Norway.

45.  Henriksen, A.  and D. F. Brakke. 1986.
     Estimates of  Critical  Loads  for Sulfur  to
     Surface Waters. Environ. Sci. Technol.  (in
     review).

46.  Hidy, G.  M. and J. R. Young.  1986.  Acid
     Deposition  and  the  West, a  Scientific
     Assessment.  ERT  Document No.  P-D572-
     503, ERT Inc., Newbury Park, California. 187
     pp.

47.  Hillman, D. C., J. F. Potter, and S. J. Simon.
     1986.  National Surface  Water  Survey,
     Eastern Lake Survey - Phase I, Analytical
     Methods Manual. Environmental Monitoring
     Systems  Laboratory, U.S. Environmental
     Protection Agency, Las Vegas, Nevada.

48.  Hoffman,  G. R. and  R.  Alexander.  1976.
     Forest Vegetation of the Bighorn Mountains,
     Wyoming: a Habitat  Classification. USDA
     Forest Service Research Paper RM-170. 38
     pp.

49.  Hunt,  C.  B. 1974.  Natural Regions of the
     United States and Canada. W.H. Freeman,
     San Francisco, California. 725 pp.

50.  Kanciruk, P., J. M. Eilers, R. A. McCord,  D.
     H. Landers, D. F. Brakke, and R. A. Linthurst.
     1986a. Characteristics  of Lakes  in the
     Eastern  United States.  Volume  III:  Data
     Compendium  of  Site Characteristics and
     Chemical  Variables. EPA/600/4-86/007c,
     U.S.  Environmental Protection  Agency,
     Washington, D.C. 374 pp.

51.  Kanciruk, P., M. J. Gentry, R. A. McCord, L.
     A. Hook, J. M. Eilers, and M. D. Best. 1986b.
     National Surface Water Survey, Eastern Lake
     Survey -  Phase I, Data Base Dictionary. Oak
     Ridge National Laboratory, TM-10153, Oak
     Ridge  National  Laboratory, Oak Ridge,
     Tennessee. 85 pp.

52.  Kanciruk, P., R. J. Olson, and R. A. McCord.
     1986c.   Data  Quality  Assurance  in  a
     Research  Data Base: The  1984  National
     Surface  Water Survey,  pp.  193-207.  In:
     Research Data Management  in the Ecolog-
     ical Sciences,  Belle W.  Baruch Library of
     Marine Science No. 16, W. K. Michener, ed.
     University of South Carolina  Press, Colum-
     bia, South Carolina.

53.  Kerfoot, H. B. and M. L. Faber. 1987. National
     Surface  Water Survey, Western  Lake
     Survey -  Phase I, Analytical Methods  Man-
     ual.  Environmental  Monitoring  Systems
     Laboratory, U.S. Environmental Protection
     Agency, Las Vegas, Nevada.

54.  Kirchmer, C. J.  1983.  Quality Control in
     Water Analyses. Env. Sci. Tech.  17:174A-
     181A.

55.  Kling, G. W. and M. C.  Grant. 1984. Acid
     Precipitation in the Colorado Front Range:
     An Overview  with Time Predictions for
     Significant Effects. Arctic and  Alpine
     Research 16:321-329.

56.  Kohler, M. A., T. J. Nordenson, and  D. R.
     Baker. 1955.  Evaporation Maps  for the
                  152

-------
     United  States.  U.S.  Weather  Bureau
     Research Paper 38. 21 pp.

57.  Krug, E. C. and C. R. Frink.  1983. Acid Rain
     on Acid Soil:  A New Perspective.  Science
     221:520-525.

58.  Laird, L. B., H. E. Taylor, and V. C. Kennedy.
     1986.  Snow  Chemistry of the Cascade-
     Sierra Nevada Mountains. Env. Sci. Technol.
     20:275-290.

59.  Larson,  E. S.  and R. G.  Schmidt.  1958. A
     Reconnaissance of  the Idaho Batholith and
     Comparison with the  Southern California
     Batholith.  U.S. Geological  Survey  Bulletin
     1070-A. 33pp.

60.  Lewis, W. M., Jr. 1982. Changes in pH and
     Buffering Capacity of Lakes in the Colorado
     Rockies. Limnol. Oceanogr.  27:167-172.

61.  Lewis,  W. M., Jr. and M.  C. Grant.  1979.
     Acid Precipitation  in the  Western United
     States. Science 207:176-177.

62.  Lewis, W. M., M. C. Grant, J. F. Saunders.
     1984. Chemical  Patterns  of Bulk Atmos-
     pheric Deposition in the State of Colorado.
     Water Resources Research  20:1691-1704.

63.  Likens, G. E. 1 976. Acid Precipitation. Chem.
     Eng. News 54:29-44.

64.  Linthurst, R. A., D.  H. Landers, J. M.  Eilers,
     D. F. Brakke, W. S.  Overton, E. P.Meier, and
     R. E. Crowe. 1986.  Characteristics  of Lakes
     in the  Eastern  United  States.  Volume  I:
     Population  Descriptions  and  Physico-
     Chemical  Relationships. EPA/600/4-86/
     007a,   U.S.  Environmental  Protection
     Agency, Washington, D.C. 136 pp.

65.  Loranger, T. J. 1986. Temporal Variability of
     Water and Snowpack Chemistry in the North
     Cascade Mountains. M. S.  Thesis,  Western
     Washington   University,  Bellingham,
     Washington. 140 pp.

66.  Loranger,  T.  J. and D. F. Brakke.  1986.
     Temporal Variability of  Lake Waters  in the
     North Cascades  Mountains (Washington,
     USA). Water, Air, and Soil Poll. 30:123-129.

67.  McColl, J. G. 1981.  Increasing Hydrogen Ion
     Activity of Water in Two Reservoirs  Supp-
     lying the San Francisco Bay Area, California.
     Water Resources Research  17:1510-1516.

68.  McFee,  W. W.  1980.  Sensitivity of  Soil
     Regions to Acid Precipitation. EPA/600/3-
     80/013. Corvallis Environmental Research
     Laboratory, U.S. Environmental Protection
     Agency, Corvallis, Oregon. 178 pp.

69.  Meglen, R. R., W. R. Chappell, R. B. McNelly,
     R.  J. Sistko, G. A.  Swanson, L. A. Taylor,
     and R.  W. Klusman. 1985. Chemical  Clas-
     sification of Colorado Lakes,  Part I. Center
     for Environmental  Sciences. University of
     Colorado at Denver, Colorado. 224 pp.

70.  Melack, J. M.,  J.  L. Stoddard,  and  D. R.
     Dawson. 1982. Acid Precipitation and Buffer
     Capacity  of  Lakes  in  the Sierra  Nevada,
     California, pp  35-41.  In:  International
     Symposium on Hydrometerology, American
     Water Resources Association. June 13-17,
     1982, Denver, Colorado.

71.  Melack, J. M., J. L.  Stoddard,  D. R. Dawson,
     and C. A. Ochs. 1 983. Assessment of Acidity
     of  Lakes and  Precipitation  in  the Sierra
     Nevada. Technical Completion Report, Pro-
     ject UCAL- WRC-W-590. California Water
     Resources Center,  University of California,
     Davis, California. 36 pp.

72.  Melack, J. M., J. L. Stoddard, and  C. A. Ochs.
     1985. Major Ion Chemistry and  Sensitivity
     to Acid  Precipitation of Sierra  Nevada Lakes.
     Water Resources Research 21:27-32.

73.  Metcalf, R. C. 1984. Field pH Determination
     in Glacial Melt Waters. J.  Glaciology 30:106-
     111.

74.  Morris, F. A., D.  V. Peck, D.  C. Hillman, K.
     J. Cabbie, and S. L. Pierett. 1986. National
     Surface Water Survey, Western Lake Survey
     -  Phase I, Field Training  and Operations
     Manual.  Internal  Report. Environmental
     Monitoring Systems Laboratory,  U.S. Envir-
     onmental Protection  Agency, Las  Vegas,
     Nevada. 177 pp.

75.  Munger, J. W. and F.  G. Eisenreich.  1983.
     Continental Scale Variations in Precipitation
     Chemistry. Env. Sci. Tech. 17:32A-42A.

76.  National Acid Deposition Program (NADP).
     1984.  Unpublished annual summaries.
     Natural Resource Ecology Lab,  Ft.  Collins,
     Colorado.

77.  Nelson, P.  O. 1985. Acid Rain Buffering
     Potential in Oregon Cascade  Lakes: Secon-
     dary  Mineral  Solubility Control of Solution
     Ionic  Composition.  Water  Resources
     Research Institute,  Oregon State University,
     Corvallis. Oregon. 16 pp.
                                                                       153

-------
78.  Nelson,  P. 0. and  G.  K.  Delwiche.  1983.
     Sensitivity of Oregon's Cascade  Lakes to
     Acid Precipitation. Project Completion Report
     A-061-ORE. Water Resources Research
     Institute, Oregon State University,  Corvallis,
     Oregon.  53 pp.

79.  Norton, S.  A., J. J. Akielaszek, T. A. Haines,
     K. L. Stromborg, and J. R. Langcone. 1982.
     Bedrock  Geologic  Control of Sensitivity of
     Aquatic  Ecosystems in the United  States to
     Acidic Deposition. 11  pp.  Explanatory Text
     for National Atmospheric Deposition Pro-
     gram.  Distribution of  Surface Waters Sen-
     sitive to  Acidic Precipitation: A State-Level
     Atlas. University  of  Maine,  Geological
     Sciences, Orono, Maine. 66 pp.

80.  Norton,  S. A. and A.  Henriksen. 1983. The
     importance of C02 in  Evaluation  of Effects
     of Acidic Deposition. Vatten 39:346-354.

81.  Ogden,  J. G.,  III. 1982. Seasonal  Mass
     Balance of Major Ions in Three  Small
     Watersheds in a  Maritime Environment.
     Water, Air, and Soil Poll. 17:119-1 30.

82.  Oliver,  B. G.,  E.  M.  Thurman, and  R.  L.
     Malcolm.  1983. The Contribution  of Humic
     Substances to the Acidity of  Colored Natural
     Waters.  Geochim.  Cosmochim.  Acta
     47:2031-2035.

83.  Omernik, J. M. and G. E. Griffith. 1986. Total
     Alkalinity  of Surface  Waters: a Map of the
     Western Region. EPA/600/D-85/219. Cor-
     vallis  Environmental Research Laboratory,
     U.S.  Environmental  Protection  Agency,
     Corvallis, Oregon. 52 pp.

84.  Omernik, J. M. and C. F. Powers. 1983. Total
     Alkalinity of Surface Waters—a National
     Map. Annals of the Association of American
     Geographers 73:133-1 36.

85.  Oppenheimer, M., C.  B. Epstein,  and R. E.
     Yuhnke. 1985.  Acid  Deposition, Smelter
     Emissions, and the Linearity Issue in the
     Western United States. Science 229:859-
     861.

 86.  Overton, W. S. and D. J. Blick. 1986. Effects
     of Measurement  and Other Extraneous
     Errors  on  Estimated  Cumulative Distribu-
     tions  in the National Lake Survey.  Presented
     at the  Western Regional  Meeting  of the
     Institute  of Mathematical  Statistics,  July
     1986. (Unpublished).
87.  Overton, W. S.,  P. Kanciruk, L. A.  Hook, J.
     M. Eilers, D. H. Landers, D. F.  Brakke, D. J.
     Blick, Jr., R. A. Linthurst, and M. D. DeHaan.
     1986. Characteristics of Lakes in the Eastern
     United States. Volume II: Lakes Sampled and
     Descriptive  Statistics for Physical  and
     Chemical Variables.  EPA/600/4-86/007b,
     U.S.  Environmental Protection Agency,
     Washington, D.C. 369 pp.

88.  Peck, D. V., R. F. Cusimano, and W. L. Kinney.
     1985. National  Surface  Water  Survey,
     Western Lake  Survey  - Phase I,  Ground
     Sampling Training and Operations  Manual.
     Internal  draft.  Environmental Monitoring
     Systems  Laboratory. U.S. Environmental
     Protection  Agency, Las Vegas, Nevada. 37
     pp.

89.  Pennak, R. W. 1963. Rocky Mountain States,
     pp. 349-369. In: Limnology in North America,
     D. G. Frey, ed. University of Wisconsin Press,
     Madison, Wisconsin.

90.  Permutt, T. and M. Moezzi. 1986.  Relative
     Interlaboratory  Bias  in  the Western Lake
     Survey.  Systems Applications, Inc.  San
     Rafael, California. 43 pp.

91.  Rasmussen, L. A. and W. V. Tangborn. 1976.
     Hydrology  of the  North Cascades  Region,
     Washington. I.  Runoff, Precipitation  and
     Storage  Characteristics. Water  Resources
     Research 12:187-202.

92.  Reddy,  M. M. and H. C. Claassen.  1985.
     Estimates of Average Major Ion Concentra-
     tions in Bulk Precipitation at Two High-
     Altitude  Sites Near the Continental  Divide
     in  Southwestern Colorado.  Atmospheric
     Environment 19:1199-1203.

93.  Reynolds,  R.  C.  and  N.  M. Johnson. 1972.
     Chemical  Weathering  in  the Temperate
     Glacial  Environment of the Northern Cas-
     cade Mountains. Geochim.   Cosmochim.
     Acta. 36:537-554.

94.  Robertson, C.  E.  and  J.  D.  Hem.  1969.
     Solubility of Aluminum in the Presence of
     Hydroxide, Fluoride and Sulfate. U.S. Geo-
     logical Survey Water-Supply Paper  1827-C.
     37pp.

95.  Rodhe,  W. 1949. The Ionic Composition of
     Lake Waters. Verh.  Int. Verein.  Limnol.
     13:121-141.

96.  Rosen,  A. E. and P.   Kanciruk. 1985. A
     Generic  Data Entry Quality Assurance Tool.
                   154

-------
      Proceedings of the  Tenth  Annual Interna-
      tional SAS (Statistical Analysis System, Inc.)
      Users  Group  Conference, March 10-13,
      1985. Reno, Nevada.

  97.  Roth, P., C. Blanchard, J. Harte, H. Michaels,
      and  M. T. EI-Ashry. 1985. The  American
      West's Acid Rain Test. Research Report #1,
      World Resources Institute, Washington, D.C.
      50pp.

  98.  SAS Institute. 1982. The Statistical Analysis
      System, SAS Institute, Gary, North Carolina.

  99.  SAS Institute.  1985.  SAS/GRAPH  User's
      Guide, Version  5 Edition, Appendix 2,  SAS
      Institute, Gary, North Carolina. 596 pp.

 100.  Shilts,  W. W. 1981. Sensitivity of Bedrock
      to Acid Precipitation; Modification by Glacial
      Processes. Geol. Surv. Can. Paper 81-14.

 101.  Silverstein, M. E., S. K. Drouse, J. L. Engels,
      M. L. Faber, and T. E. Mitchell-Hall. 1987a.
      National  Surface  Water  Survey,  Western
      Lake Survey -  Phase I,  Quality Assurance
      Plan. EPA/600/8-87/026. Environmental
      Monitoring Systems Laboratory, U.S. Envir-
      onmental  Protection Agency,  Las Vegas,
      Nevada.

 102.  Silverstein, M.  E., M. L. Faber, S. K. Drouse,
      and  T.  E.  Mitchell-Hall.  1987b.  National
      Surface  Water Survey,  Western  Lake
      Survey - Phase I, Quality Assurance Report.
      Environmental Monitoring  Systems Labora-
      tory, U.S. Environmental Protection Agency,
      Las Vegas, Nevada.

103.  Smith,  R.  A. and R. B. Alexander.  1983.
      Evidence  for Acid  Precipitation  Induced
      Trends  in  Stream  Chemistry at Hydrologic
      Bench-Mark  Stations. U.S. Geological Sur-
      vey Circular 910. 12  pp.

104.  Snead,  R. E.  1980. World Atlas of Geomor-
      phic  Features. Robert E. Krieger,  Inc.,
      Huntington, New York, New York. 301 pp.

105.  Steele,  R., S. Cooper, D. Ondov, D. Roberts,
      and R.  Pfister.  1983. Forest Habitat  Types
      of Eastern  Idaho  and Western  Wyoming.
      USDA  Forest Service  General  Technical
      Report. INT-144. Intermountain Forest  and
      Range Experiment Station, Ogden, Utah.  1 22
      PP-

106.  Stensland, G. J. 1984. Chapter 8: Deposition
      Monitoring, pp.  8-32. In: The Acidic  Depo-
      sition Phenomenon and Its Effects. Critical
      Assessment Review  Papers,  Volume I:
      Atmospheric Sciences, A. P. Altshuller and
      R. A. Linthurst eds., EPA/600/8-83/016 AF,
      U.S.  Environmental  Protection  Agency,
      Washington, D.C.

 107.  Stensland, G. J., D. M. Whelpdale, and G.
      Oehlert.  1986.  Chapter 5.  Precipitation
      Chemistry, pp. 128-199. In: Acid Deposition,
      Long Term Trends. National Academy Press,
      Washington, DC. 506 pp.
108.  Stuart, S. A. 1984. Hydrology and Aquatic
      Chemistry of Monitor Lake Watersheds in the
      Wind River Mountains, Wyoming, pp. 130-
      173. In: Workshop Proceedings - Air Quality
      and  Acid Deposition  in  the Bridger and
      Fitzpatrick  Wildernesses.  USDA Forest
      Service, Intermountain Region, Ogden, Utah.
      292 pp.

109.  Stumm, W. and J. J. Morgan. 1 981. Aquatic
      Chemistry: An Introduction  Emphasizing
      Chemical Equilibria in Natural Waters. 2nd
      Edition. John Wiley & Sons, New York, New
      York. 779 pp.

110.  Sumioka, S. S. and  N. P. Dion. 1 985. Trophic
      Classification of Washington Lakes  Using
      Reconnaissance Data. Washington  State
      Department of Ecology, Water Supply Bul-
      letin 57.  320pp.

111.  Swedish  Ministry  of Agriculture. 1982.
      Acidification Today and Tomorrow. Ministry
      of Agriculture, Environment Committee,
      Stockholm, Sweden. 232 pp.
112.  Tonnessen, K. A. 1983. Potential for Aquatic
      Ecosystem Acidification in the  Sierra Nev-
      ada,  147-169 pp.  In:  Acid Precipitation:
      Aquatic Effects, G. Hendrey, ed., Ann Arbor
      Science, Woodburn, Masachusetts. 173 pp.

113.  Tukey, J. W.  1977. Exploratory Data Anal-
      ysis.  Addison-Wesley,  Reading, Massachu-
      setts. 688 pp.

114.  Turk, J.  T.  1983. An Evaluation of Trends
      in the Acidity of Precipitation and the Related
      Acidification of Surface Waters  in  North
      America. U.S. Geological  Survey  Water-
      Supply Paper 2249. 18 pp.

115.  Turk, J. T. and D. B. Adams. 1983. Sensitivity
      to  Acidification of Lakes  in the Flat Tops
      Wilderness Area, Colorado. Water Resour-
      ces Research 19:346-350.
                                                                        755

-------
116.  U.S. Department of Agriculture-Soil Conser-
      vation Service.  1976. Scale 1:1,000,000.
      Climax Vegetation. Montana.

117.  U.S.  Geological  Survey.  1970.  Potential
      Natural Vegetation. In: National Atlas of the
      United States of America. Washington, D.C.
      417pp.

118.  U.S.  Department of Commerce.  1974.  Cli-
      matic Atlas of the United States. 80 pp.

119.  U.S.  Environmental  Protection Agency.
      1983. Methods for Chemical Analysis of
      Water and  Wastes.  EPA/600/4-79/020,
      Environmental  Monitoring  and Support
      Laboratory, Cincinnati, Ohio.

120.  Van Alstine,  R. E. and E. W. looker. 1979.
      Preliminary Map of Fluorine Provinces in the
      Conterminous  United States. In: Atlas of
      Metal and Nonmetal Mineral Provinces of the
      Conterminous United States.  United States
      Geological Survey Open-File Report  79-
      576G.

121.  Velleman,  P. F. and  D. C. Hoaglin. 1981.
      Applications,  Basics and Computing of
      Exploratory  Data Analysis. Duxbury Press,
      Boston, Massachusetts.

122.  Welch, E. B., D. E. Spyridakis, and T. Smayda.
      1986. Acidification  Potential in High Moun-
      tain Lakes in Washington: 3. Final  Report,
      Washington State Department of Ecology. 24
      PP.

123.  Wetzel, R. G. 1983. Limnology. 2nd Edition.
      Saunders College Publishing, Philadelphia,
      Pennsylvania. 767 pp.

124.  Wilderness Act of 1964. Public Law 88-577.

125.  Wilson,  D. and R.  Wood. 1984.  Mountain
      Lake Water Quality Survey. 1983. California
      Department  of Fish and Game,  Region 1,
      Data Report.  7 pp.

126.  Wissmar, R.  G.,  J.  E. Richey, A. H. Devol,
      and D. M. Eggers.  1982.  Lake Ecosystems
      of the Lake Washington Drainage Basin, pp.
      333-385. In: Analysis of Coniferous Forest
      Ecosystems in  the  Western United States,
      R. L. Edmonds, ed. Hutchison Ross Publish-
      ing  Company,  Stroudsburg,  Pennsylvania.
      419pp.
                   156

-------
                                         Section  11
                                           Glossary
11.1
A
AAS
Ac
ANC
APHA

ASTM
        Abbreviations
              —lake surface area contained in a
                population of target lakes
              —estimated total lake surface area
                in a target population
              —atomic absorption spectroscopy
              —estimated lake surface area
                having a value for a variable
                below Xc
              —95 percent upper confidence limit
                on Ac
              —acid neutralizing capacity
              —American Public Health
                Association
              —American Society for Testing and
                Materials
CaCOa        —calcium carbonate or calcite
CaMg (C03)2   —dolomite
CaSC>4        —anhydrite or anhydrous calcium
                sulfate
CaSO4-2H2O   —gypsum
cm            —centimeter
DDRP         —Direct/Delayed Response Project
DGC          —duplicate (ground crew)
DHC          —duplicate (helicopter crew)
DIG           —dissolved inorganic carbon
DOC          —dissolved organic carbon
DQO          —data quality objective
ELS-I          —Eastern Lake Survey - Phase I
EMSI          —Environmental Monitoring and
                Services, Inc.
EMSL-
Las Vegas     —Environmental Monitoring
                Systems Laboratory - Las Vegas
EPA           —Environmental Protection Agency
ERL-Corvallis  —Environmental Research
                Laboratory - Corvallis
ERP           —Episodic Response Project
ExtAI         —extractable aluminum
FL            —field low synthetic audit
FN            —field natural audit
ft             —feet
F(x)           —cumulative frequency distribution
gc            —reference proportion (lake area)
G(x)           —cumulative areal distribution
ha            —hectare
IBM           —International Business Machines
ICPAES       —inductively coupled plasma
                atomic emission spectroscopy
ID            —identification number
IR            —infrared
km2           —square kilometers
L             —liter
LA            —lake area
Lockheed-     —Lockheed Engineering and
EMSCO         Management Services Company,
                Inc.
m            —meter
M            —median value of an F(x) or G(x)
                distribution
MAX          —maximum value of variable X
mg L"1        —milligrams per liter
mi2           —square miles
MIBK-HQ     —methyl isobutyl ketone-
                hydroxyquinoline
                                                 MIN
                                                 mL
                                                 mm
                                                 N

                                                 N
               -minimum value of variable X
               -miHiliter
               -millimeter
               -normal (normality), as in
                equivalents per liter
               -number of lakes in a target
                population
                                              757

-------
N

N*

Nc
Nc

Ncu

Nn

n*
n*
n'
nc

nna
n0
nz

NADP/NTN

NAPAP

NBS
NLS
NSWS
NTU
ORNL
P
PC

PCU
ppm
PRECIP
—estimated number of lakes in a
  target population
—number of lakes in map
  population
—number of lakes below Xc
—estimated number of lakes below
  Xc
—95 percent upper confidence limit
  on No
—estimated nontarget population
  size
—number of lakes selected in
  probability sample
—number of lakes visited
—number of target lakes sampled
—effective sample size
—number of lakes in the probability
  sample below Xc
—number of lakes in the probability
  sample determined after visiting
  the lake
—number of non-target lakes in the
  probability sample determined
  from maps before visitation
—number of lakes not visited
—number of sample lakes in
  subpopulation z
—National Atmospheric Deposition
  Program/National Trends
  Network
—National Acid Precipitation
  Assessment Program
—National Bureau of Standards
—National Lake Survey
—National Surface Water  Survey
—nephelometric turbidity  unit
—Oak Ridge National Laboratory
—inclusion probability
—reference proportion (number of
  lakes)
—platinum cobalt unit
—parts per million
—variable name for precipitation
—proportion of lakes visited
—first quintile of F(x) or G(x); 20th
  percentile
Q3

CU
QA
QC
QCCS
Qd
RDL
RGC
RHC
RMS
RO
%RSD

RT
SAS
SB

SO
SDL
SE
STP

SIR
THC
UCL
fjtm
fjS cm"
USGS
UV
V
W
WA
WLS-I
WMP
x
—second quintile of F(x) or G(x);
  40th percentile
—third quintile of F(x) or G(x); 60th
  percentile
—fourth quintile of F(x) or G(x); 80th
  percentile
—fifth quintile of F(x) or G(x);
  maximum
—quality assurance
—quality control
—quality control check sample
—interquintile difference
—required detection limit
—routine (ground crew)
—routine (helicopter crew)
—root mean square
—runoff
—percent relative standard
  deviation
—residence time
—Statistical Analysis System
—standard  deviation of field blank
  measurements
—standard  deviation
—system decision limit
—standard  error
—standard  temperature and
  pressure
—stratum
—triplicate  (helicopter crew)
—95 percent upper confidence limit
—microequivalents per liter
—micrograms per liter
—micrometers
—microsiemens per centimeter
—United States Geological Survey
—ultraviolet
—estimated variance
—weight or expansion factor
—watershed area
—Western  Lake Survey - Phase I
—Watershed Manipulation Project
—measured value of variable
                     158

-------
X

x

Xc

YSI

z

Zs
—variable
—mean value for a variable x
—reference value of the variable X
—Yellow Springs Instruments
—subpopulation of lakes
—site depth
11.2   Definitions

ABSOLUTE BIAS—the expected difference between
  a measured value and the true value.

ACID  NEUTRALIZING  CAPACITY—the total acid-
  combining capacity of a water sample determined
  by titration with a strong acid. Acid neutralizing
  capacity includes ALKALINITY (carbonate species)
  as  well as other basic  species (e.g.,  borates,
  dissociated  ORGANIC  ACIDS, alumino-hydroxy
  complexes).

ACIDIC  DEPOSITION—rain, snow, or dry fallout
  containing  high concentrations of sulfuric acid,
  nitric acid or hydrochloric acid, usually produced
  by atmospheric transformation of the by-products
  of fossil fuel combustion (power plants, smelters,
  autos, etc.); precipitation with  a pH less than 5.0
  is generally considered to be unnaturally acidic.

ACIDIC  LAKE—for  this report,  a lake with ACID
  NEUTRALIZING CAPACITY less than or equal to
  0/ueq L"1.

ACIDIFICATION—any temporary or permanent loss
  of ACID NEUTRALIZING CAPACITY.

ACCURACY—the closeness of a measured value to
  the true value of an ANALYTE.

AIR-EQUILIBRATED—a sample ALIQUOT that has
  been brought to equilibrium with standard air (300
  ppm C02> before analysis; used with some pH and
  DISSOLVED INORGANIC CARBON  measure-
  ments.

ALGORITHM—a step-by-step mathematical proce-
  dure for solving a problem.

ALIQUOT—portion of a sample  prepared for  the
  analysis of particular  constituents,  sent in  a
  separate container to the ANALYTICAL LABOR-
  ATORY.

ALKALINITY—the  titratable base  of  a sample
  containing  hydroxide, carbonate, and bicarbonate
  ions,  i.e.,  the  equivalents of  acid required  to
  neutralize the basic carbonate components.
ALKALINITY  MAP CLASS—a geographic  area
  defined by the expected ALKALINITY of a majority
  of SURFACE WATERS (does not necessarily reflect
  measured alkalinity); used as a STRATIFICATION
  FACTOR in  the  lake  selection  design. Three
  alkalinity map classes were used: < 100, 100-199,
  and 200-400 ,ueq L"1.

ALLUVIAL—deposited by rivers and streams.

AMONG-BATCH PRECISION—the estimate of  PRE-
  CISION that  includes effects of different labora-
  tories and day-to-day difference within a single
  laboratory.

ANALYTICAL  LABORATORY—in  this  report,  a
  laboratory under contract with the U.S. Environ-
  mental Protection Agency to analyze water
  samples shipped from the FIELD LABORATORIES.

ANALYTE—a chemical species that is measured in
  a water sample.

ANDESITE—dark  gray  volcanic  rock  composed
  largely of FELDSPAR, often  having large feldspar
  crystals in a finer matrix.

ANION—a negatively charged  ion.

ANION-CATION BALANCE—in an  electrically neu-
  tral solution, such as water, the total  charge of
  positive ions (CATIONS) equals the total charge
  of negative ions (ANIONS).

ANION  DEFICIT—the concentration in /ueq L"1 of
  measured ANIONS less than measured CATIONS.

ANTICLINE—an arching fold in layered rocks.

ANTHROPOGENIC—from a human source.

ASTM TYPE I REAGENT GRADE WATER—deionized
  water used for BLANK samples and REAGENT
  preparation that had a measured CONDUCTANCE
  less than 1  pS cirT1.

AUDIT—an on-site evaluation of field sampling or
  laboratory activities to  verify that standardized
  protocols are being followed.

AUDIT   SAMPLE—a  standardized  water sample
  submitted to laboratories to check overall perfor-
  mance in sample analysis. Natural audit samples
  were  lake water; synthetic audit samples  were
  prepared  by diluting  concentrates  of known
  chemical composition. See FIELD AUDIT SAMPLE
  and LABORATORY AUDIT SAMPLE.

BASALT—dark gray to black volcanic rock which is
  poor  in silica and rich  in iron and  magnesium
  minerals.
                                                                       755

-------
BASE CATION—a non-protolytic cation that does not
  affect ACID NEUTRALIZING CAPACITY;  usually
  calcium or magnesium. See PROTOLYTE.

BATCH—a group of samples processed and analyzed
  together. A field batch of samples is defined as
  all  samples (including QUALITY ASSURANCE
  samples) processed at one FIELD LABORATORY
  in one day. A laboratory batch is defined as all
  samples processed and analyzed at one ANALYT-
  ICAL LABORATORY, associated with one set of
  laboratory QUALITY CONTROL samples.

BATHOLITH—a  large mass of  INTRUSIVE  rock,
  usually granite, more than 100 square kilometers
  (40 square miles) in surface exposure.

BEDROCK—solid rock exposed at or near the surface.

BIAS—the  systematic difference  between the
  measured and the  true values. See ABSOLUTE
  BIAS and RELATIVE BIAS.

BLANK—a sample of ASTM TYPE I REAGENT GRADE
  WATER analyzed as  a  QUALITY ASSURANCE/
  QUALITY  CONTROL sample in the  WLS-I. See
  FIELD  BLANK SAMPLE,  LABORATORY  BLANK
  SAMPLE, and TRAILER BLANK.

BUFFERING CAPACITY—the quantity of acid or base
  that can be added to a water sample with little
  change in pH.

CALCIUM BICARBONATE  SYSTEM—see CARBO-
  NATE SYSTEM.

CALCULATED  CONDUCTANCE—the  sum of the
  theoretical specific conductances of all measured
  ions in a sample.

CALCAREOUS—made  of  or containing calcium
  carbonate. See CALCITE.

CALCITE—calcium  carbonate (CaCOs); a common
  rock-forming  mineral, the principal  mineral in
  limestone, marble, and chalk.

CALCIUM—a silver-white metallic element of the
  alkaline-earth  group occurring only in compound
  forms.

CALC-SILICATE  GNEISS—a  METAMORPHIC rock
  consisting of mainly calcium-bearing  silicate
  minerals.

CALDERA—a broad, basin-shaped volcanic depres-
  sion formed  by the explosion or collapse of a
  magma chamber.

CALIBRATION BLANK—a  solution used in calibrat-
  ing or standardizing analytical instruments.
CALIBRATION STUDY—a study  conducted during
  the WLS-I to determine if the method of sample
  collection (helicopter crew  vs. ground crew)
  affected  the chemistry of the water samples;
  samples collected during this study were also used
  to evaluate analytical laboratory bias.

CARBONATE SYSTEM—a lake in which the major
  part of ACID NEUTRALIZING  CAPACITY is  com-
  posed of  carbonate; organic or other weak ahions
  contribute less than 10 percent to the total ANION
  charge.

CATCHMENT—see WATERSHED.

CATION—a positively charged ion.

CATION EXCHANGE—a reversible process occurring
  in soil  in which acidic CATIONS (e.g., hydrogen
  ions) are adsorbed and BASE CATIONS are
  released.

CHELATOR—a class of  compounds, organic  and
  inorganic, that can bind metal ions  and  change
  their biological availability.

CIRCUMNEUTRAL—close to neutrality with respect
  to pH (pH ~ 7).

CIRQUE—a steep-walled, usually semicircular, basin
  excavated by the head of a glacier.

CLEARWATER LAKE—for this report, .a lake having
  TRUE COLOR less than or equal to 30 PLATINUM
  COBALT  UNITS.

CLOSED LAKE—a lake with a surface  water inlet
  but no surface water outlet.

CLOSED SYSTEM—method of  measurement in
  which a  water sample is collected and analyzed
  for pH and  DISSOLVED  INORGANIC CARBON
  without exposure to the atmosphere.

CLUSTER ANALYSIS—a multivariate classification
  technique for identifying similar (or dissimilar)
  groups of observations.

COMPARABILITY—a measure of data quality that
  allows the similarity within and among data sets
  to be established confidently.

COMPLETENESS— a measure of data quality that
  is the quantity of acceptable data actually collected
  relative to the total quantity that was attempted
  to be collected.

CONDUCTANCE—a measure of the electrical
  conductance (the  reciprocal  of the electrical
  resistance) or total IONIC STRENGTH of a water
  sample.
                    760

-------
CONGLOMERATES—rocks composed  of  rounded,
  waterworn fragments  of older  rocks,  cemented
  together by another mineral substance.

CRETACEOUS—the third and latest of the periods
  included in the MESOZOIC Era; also the system
  of strata deposited in the Cretaceous Period.

CUMULATIVE FREQUENCY DISTRIBUTION—a func-
  tion, F(x), such that for  any reference value X, F(x)
  is the  estimated proportion of lakes in  the
  population having a value x < X.

CUMULATIVE AREAL  DISTRIBUTION—a  function,
  G(x), such that for any REFERENCE VALUE X, G(x)
  is the estimated proportion  of lake area in the
  population having a value x < X.

DARKWATER LAKE—for this  report,  a  lake with
  TRUE COLOR greater than 30 PLATINUM COBALT
  UNITS.

DATA BASE—all computerized results of the survey,
  which include DATA SETS 1, 2, 3, and 4, as well
  as back-up and historical data sets.

DATA FILE—a subset of information, often in tabular
  form (e.g., the lake location file contains informa-
  tion on lake identification and location).

DATA PACKAGE—a report, generated by an ANA-
  LYTICAL  LABORATORY for  each BATCH of
  samples analyzed, that  included analytical results,
  ACID NEUTRALIZING CAPACITY titration data, and
  QUALITY CONTROL CHECK SAMPLE results.

DATA QUALIFIER—see FLAG and TAG.

DATA QUALITY OBJECTIVES—ACCURACY, DETEC-
  TABILITY, and PRECISION limits desired by data
  users, established prior to the beginning of a
  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, enhanced
  data;  missing values or errors in the VALIDATED
  data set were replaced by  substitution values,
  duplicate values were averaged, and  negative
  measurements (except ACID  NEUTRALIZING
  CAPACITY) were set equal to zero.

DETECTABILITY—the capacity of an instrument or
  method to  determine  a measured value for an
  ANALYTE above background  levels. See INSTRU-
  MENT DETECTION LIMIT and SYSTEM DECISION
  LIMIT.
DIKE—a tabular body of IGNEOUS rock that cuts
  across the structure of adjacent rocks  or cuts
  massive rocks. Most dikes result from the intrusion
  of  magma; some are the  result of mineral
  replacement.

DILUTE LAKE—for  this report,  a lake  with a
  CONDUCTANCE less than  10 /uS cm"1.


DISSOLVED  INORGANIC CARBON—a measure of
  the  dissolved carbon  dioxide,  carbonic acid,
  bicarbonate and carbonate anions which comprise
  the major part of ACID NEUTRALIZING CAPACITY
  in a lake.

DISSOLVED ORGANIC  CARBON—the organic
  fraction  of carbon in a  water  sample  that is
  dissolved or unfilterable (for this report, 0.45 ^m
  pore size).

DOLOMITE—a common rock-forming mineral, CaMg
  (C03>2, often with some iron replacing magnesium;
  those rocks that approximate the mineral dolomite
  in composition.

DOLOSTONE—SEDIMENTARY  rock  composed  of
  fragmental, aggregated, or precipitated DOLOMITE
  of organic or inorganic origin.

DRAINAGE BASIN—see WATERSHED.

DRAINAGE  LAKE—a  lake  with  surface  water
  outlet(s) or  with both inlets and outlets.

DUNITE—an  ULTRAMAFIC rock consisting primarily
  of olivine, (Mg, Fe>2 Si04, a common rock-forming
  mineral of basic and low-silica IGNEOUS rocks.

DUPLICATE LAKE SAMPLE—see FIELD DUPLICATE
  SAMPLE.
EFFECTIVE SAMPLE SIZE (n')—within each STRA-
  TUM, the total sample size  modified because of
  incomplete visitation.

ELECTRONEUTRALITY—having no electric charge.

EMPIRICAL DISTRIBUTION—observed  distribution
  of results from the Survey.

EMPIRICAL RELATIONSHIP—observed relationship
  between variables based on survey results.

EQUIVALENT—unit of ionic charge; the quantity of
  a substance that either gains or loses one mole
  of protons or electrons.

EVAPOTRANSPIRATION—loss  of water from the soil
  by evaporation and by transpiration from the plants
  growing thereon.

                      161

-------
EXPANSION FACTOR—see WEIGHT.

EXTRACTABLE ALUMINUM—operationally defined
  aluminum  fraction that  is extracted by the
  procedure used in the WLS-I; this  measurement
  is intended  to  provide  an indication of the
  concentration of  aluminum that may be available
  in a form toxic to fish.

EXTRUSIVE ROCK—rock formed of magma which
  reaches  the  surface and  solidifies (also called
  volcanic rock).

FALL CIRCULATION—see TURNOVER.

FELDSPAR—a group of common, light-colored, rock-
  forming minerals containing aluminum oxides and
  silica.  Feldspars constitute  60  percent  of the
  earth's crust.

FELSIC—a mnemonic term  derived from (fe) for
  feldspar, (I) for lenads or feldspathoids, and (s) for
  silica and applied to light-colored rocks containing
  an abundance of one or all of these constituents.
  FELDSPAR  and  quartz  are  examples  of  felsic
  minerals.
FIELD  AUDIT SAMPLE—a standardized water
  sample submitted to  FIELD  LABORATORIES to
  check overall performance in sample analysis by
  both  FIELD and  ANALYTICAL LABORATORIES.
  Natural  field  audit samples  were lake water;
  synthetic field  audit samples were prepared by
  diluting concentrates of known chemical compo-
  sition.

FIELD BLANK SAMPLE—a  sample of ASTM TYPE
  I REAGENT GRADE WATER prepared at the  lake
  site by field crews; these  samples were analyzed
  by both FIELD and ANALYTICAL LABORATORIES.
FLUORITE—a clear to translucent mineral, CaF2; the
  ore of fluorine.

FLUORAPATITE—a  mineral of the apatite group,
  Cas(P04)sF; common accessory mineral in IGNE-
  OUS rocks.

FLUVIAL—of, relating to, or living in a stream or river;
  produced by stream action.

FLUVIAL  DEPOSITS—deposits produced by  the
  action of a stream or river.

FRAME—a structural representation of a population
  providing a sampling capability.

GEOMORPHIC—of or relating to the form of the solid
  surface features of the earth.
GEOMORPHIC UNIT—an area with similar physio-
  graphic or geologic characteristics. See PHYSIO-
  GRAPHY.

GIBBSITE—a mineral, Al (OH)3; a principal constit-
  uent of many bauxites (aluminum ores).

GLACIAL CIRQUE—see CIRQUE.

GLACIAL TILL—glacially deposited sediment, which
  consists  of debris of all sizes  mixed  indiscrimi-
  nately together.

GLASSY VOLCANIC ROCK (VOLCANIC GLASS)—
  natural glass produced  by the cooling of molten
  lava,  or some  liquid fraction of molten lava, too
  rapidly to  permit  crystallization, and forming
  materials such as obsidian and pitchstone.

GNEISS—banded METAMORPHIC rock thought to
  form from granite (which it commonly resembles),
  sandstone, and other continental rocks.
FIELD DUPLICATE SAMPLE—a second sample of
  lake water collected by the same crew at the same
  lake site immediately after the ROUTINE SAMPLE
  in accordance with standardized protocols.

FIELD LABORATORY—mobile laboratory located at
  each FIELD STATION in which sample processing
  and measurement of  selected  variables were
  performed.

FIELD STATION—a location providing support for
  helicopters,  sampling  personnel,  and FIELD
  LABORATORIES during field sampling operations.

FINAL DATA SET—see DATA SET 4.

FLAG—qualifier of a  data value assigned during the
  VERIFICATION and VALIDATION procedures.
GRAN  ANALYSIS—a mathematical procedure to
  determine the  equivalence points of a titration
  curve for ACID NEUTRALIZING CAPACITY.

GRANODIORITE—coarse-grained INTRUSIVE rock
  with less quartz and more FELDSPAR than granite.

GROUNDWATER—water in the part of the ground
  that is completely saturated.

GYPSUM—a common  mineral (CaSOv2H20) con-
  sisting of hydrous calcium sulfate. Alabaster is an
  example.

HALITE—a mineral form of NaCI; rock salt.

HECTARE—a measure of surface area  equal to ten
  thousand square meters or 2.47 acres.
                    762

-------
HETEROSCEDASTIC—inequality  of variances
  among groups.

HIGH  DOC LAKE—for this report,  a lake with a
  DISSOLVED ORGANIC CARBON (DOC) concentra-
  tion  > 6 mg L~1.

HIGH  ELEVATION LAKE—lakes with surface eleva-
  tion  > 3000 m above mean sea level.

HOMOSCEDASTIC—equality in variances among
  groups.

HYDRAULIC RESIDENCE TIME—the  time required
  to exchange the total volume of water in a lake.

HYDROLOGIC FLOW PATHS—the distribution and
  circulation  of water on the surface of the land,
  in the soil  and underlying rocks within a
  watershed.

IGNEOUS—rock formed from molten magma.

IN SITU—referring to measurements collected within
  the water column at a lake.

INCLUSION  PROBABILITY—the chance of a  lake
  being selected for sampling.

INDEX SAMPLE—in this report, one sample per lake,
  used to represent chemical conditions in that lake.

INITIAL DIC—a measurement of DISSOLVED INOR-
  GANIC CARBON made on an ALIQUOT  imme-
  diately before it is titrated for ACID NEUTRALIZING
  CAPACITY.

INSTRUMENTAL DETECTION LIMIT—for each chem-
  ical  variable,  a  value calculated from LABORA-
  TORY BLANK SAMPLES that  indicates the min-
  imum  concentration reliably  detectable by the
  instrument(s)  used.

INTERLABORATORY BIAS—systematic differences
  in performance between  laboratories estimated
  from analysis of the same type of AUDIT SAMPLES.

INTERMEDIATE VOLCANIC ROCKS—IGNEOUS rock
  containing between 52 and 66 percent silica (SiOa).

INTERQUARTILE RANGE—the difference between
  the  upper  and lower QUARTILE  values in a
  population frequency distribution. See QUARTILE.

INTERQUINTILE  DIFFERENCE—the  difference
  between the fourth QUINTILE (80th percentile) and
  first QUINTILE (20th  percentile)  values in a
  population frequency distribution.

INTRALABORATORY PRECISION GOAL—a preci-
  sion goal based on the DATA QUALITY OBJEC-
  TIVES for the analysis of  laboratory duplicates
  within a single laboratory.

INTRUSIVE—describing magma (naturally occurring
  molten rock material, generated within the earth)
  that has been injected into a pre-existing rock.

INVERSE CUMULATIVE AREAL DISTRIBUTION—a
  function, 1-G(x), such that for any reference value
  X, 1-G(x) is the estimated proportion of lake area
  in the population having a value x > X.

INVERSE CUMULATIVE FREQUENCY DISTRIBU-
  TION—a function,  1-F(x), such  that  for any
  reference value X, 1-F(x) is the estimated number
  of lakes in the population having a value x > X.

ION BALANCE—see ANION-CATION BALANCE.

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 concentrations of  ANIONS and
  CATIONS present in the solution.


ISOTHERMAL—defined as a temperature difference
  in a lake of less than 4°C  between the reading
  at 1.5 meters below the surface and  at 1.5 meters
  above lake bottom (synonymous with NONSTRAT-
  IFIED or mixed).

JURASSIC—the middle of three periods comprising
  the MESOZOIC Era. The system of strata deposited
  during that period.

KAOLINITE—a  clay mineral commonly  used  in
  making porcelain and fire bricks.

LABORATORY AUDIT SAMPLE—During the  ELS-I,
  a standardized water sample submitted blind to
  ANALYTICAL  LABORATORIES  to  check overall
  performance in sample analysis. All laboratory
  audit samples were synthetic, prepared by diluting
  concentrates of known chemical composition. This
  sample was not used during the WLS-I.
LABORATORY BLANK SAMPLE—a sample of ASTM
  TYPE I REAGENT GRADE WATER prepared and
  analyzed by ANALYTICAL LABORATORIES.

LABORATORY DUPLICATE SAMPLE—a split sample
  prepared at the ANALYTICAL LABORATORIES.

LAKE  ID—a unique, seven-character identification
  (ID)  code given to  each lake in  the  Survey
  designating the  REGION,  SUBREGION,  and
  ALKALINITY MAP CLASS to which  the  lake
                                                                     163

-------
  belongs; e.g., lake ID4A2-011 designates the 11th
  lake in alkalinity  map class 2 of Subregion A in
  Region 4.

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 MAP— 1:24,000-, or 1:62,500-scale
  U.S. Geological Survey topographic maps.

LIMESTONE—a  SEDIMENTARY rock consisting
  largely of calcium carbonate.

LONG-TERM ACIDIFICATION—a partial or complete
  permanent loss of ACID NEUTRALIZING CAPACITY
  from a lake.

LONG-TERM MONITORING PROGRAM—an ongo-
  ing program sponsored by the U.S. Environmental
  Protection Agency that monitors the chemistry of
  selected lakes and streams at least three times
  per year, which is designed to detect long-term
  trends in chemistry.

LORAN-C GUIDANCE SYSTEM—a navigational
  system based upon the time-delay of received 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 an ACID
  NEUTRALIZING CAPACITY (ANC)  less than  or
  equal to 50 jueq L~1.

LOW CALCIUM LAKE—for this report,  a  lake with
  a CALCIUM concentration less than  or equal to
  50 peq L'\

LOW CONDUCTANCE—for this report,  a  lake with
  a CONDUCTANCE less than 10 /uS cm"1 (synonym-
  ous with DILUTE LAKE).

LOW DOC LAKE—for this  report,  a lake  with  a
  DISSOLVED ORGANIC CARBON (DOC) concentra-
  tion less than or equal to 2 mg L"1.

LOW pH LAKE—for this  report, a lake with a pH
  less than or equal to 5.0.

MAFIC—pertaining to or composed dominantly of the
  magnesium rock-forming silicates; said of some
  IGNEOUS rocks and their constituent  minerals.

MAJOR ANION—for this report, an ANION with a
  concentration > 5 /ueq L~1.

MAJOR CATION—for this report, a CATION with a
  concentration > 5 jueq L"1.


MAP POPULATION—within  each  STRATUM,  the

                    164
  ordered list of lakes depicted on 1:100,000-scale
  U.S. Geological Survey TOPOGRAPHIC MAPS.

MATRIX—the  physical and chemical makeup of a
  sample being analyzed.

MATRIX  SPIKE—a QUALITY CONTROL sample
  analyzed at  an ANALYTICAL LABORATORY that
  was prepared by adding a known concentration
  of ANALYTE  to a sample. Matrix spike samples
  were used to determine possible chemical inter-
  ferences within a sample that might affect the
  analytical result.

MEDIAN (M)—the value of x such that F(x) or G(x)
  = 0.5; the 50th percentile.

MESOZOIC—relating  to an  era  of geologic  time
  including the interval between the PERMIAN and
  the  TERTIARY; relating to the system of rocks
  formed in this era.

METAMORPHIC—rock derived  from  pre-existing
  rocks as they are altered by heat, pressure, and
  other processes.

METASEDIMENTARY—SEDIMENTARY rock altered
  by heat,  pressure, and  other processes but still
  retaining some sedimentary characteristics.

METAVOLCANIC—volcanic  rock  altered  by heat,
  pressure, and other processes but still  retaining
  some volcanic characteristics.

MINERAL WEATHERING—dissolution of rocks and
  minerals by erosive forces.

MIXING—see TURNOVER.

MORPHOMETRY—lake shape, size, depth; physical
  description of lake basin.

NBS-TRACEABLE—describing a REAGENT or stand-
  ard,  the  accuracy of which has been  checked
  against a certified National  Bureau of Standards
  standard.

NONSTRATIFIED—see ISOTHERMAL

NON-TARGET  LAKE—a lake that either was not the
  focus of the WLS-I objectives or could  not  be
  sampled  within the constraints of the survey.

NOT VISITED LAKE—a PROBABILITY SAMPLE lake
  scheduled for sampling  that was not visited  or
  sampled during field activities; thus, its target or
  non-target  status  is unknown. See  TARGET
  POPULATION and NON-TARGET LAKE.

NOMOGRAPH—a graph reducing a mathematical
  formula to curves so that the resultant value can

-------
  be  read on the chart coordinates for any value
  assigned to the variable involved.

ON-SITE INSPECTION—inspection of the operation
  of FIELD and ANALYTICAL LABORATORIES.

OPEN SYSTEM—a  measurement  of pH  or DIS-
  SOLVED INORGANIC CARBON obtained from a
  sample that was exposed to the atmosphere during
  collection and/or measurement.

ORGANIC ACIDS—acids possessing a carboxyl
  (-COOH) group; includes fulvic and humic acids.

ORGANIC ANION—a  weakly  dissociated organic
  molecule with a negative net ionic charge.

OROGRAPHIC—associated  with or induced by the
  presence of mountains.

OUTLIERS—observations not typical of the popula-
  tion from which the sample is drawn.

OVERALL AMONG-BATCH PRECISION—estimate of
  PRECISION calculated  from  FIELD AUDIT SAM-
  PLES that include the effects of different ANA-
  LYTICAL LABORATORIES and day-to-day differen-
  ces within a single laboratory.

OVERALL WITHIN-BATCH PRECISION—an estimate
  of PRECISION calculated from FIELD DUPLICATE
  SAMPLES, LABORATORY DUPLICATE SAMPLES,
  or TRAILER DUPLICATES and reported as ROOT
  MEAN SQUARE. See WITHIN BATCH PRECISION).

PALEOZOIC—one  of  the  eras of geologic time
  comprising  the Cambrian,  Ordovician, Silurian,
  Devonian,  Carboniferous  (Mississippian and
  PENNSYLVANIAN), and PERMIAN systems; relat-
  ing to the system of rocks formed  in this era.

PENNSYLVANIAN-AGE—in the United States, the
  sixth  period of the PALEOZOIC, approximately
  equivalent to the  Upper Carboniferous period.

PERCENT ION BALANCE DIFFERENCE—a QUALITY
  ASSURANCE procedure used to  check that the
  sum of the anion EQUIVALENTS  equals the sum
  of the cation EQUIVALENTS.  See  ANION-CATION
  BALANCE.

PERCENT RELATIVE STANDARD  DEVIATION (%
  RSD)—the STANDARD DEVIATION divided by the
  mean,  multiplied by 100.

PERMIAN—the last period of the PALEOZOIC Era.

pH—the  negative logarithm of the hydrogen ion
  activity. The pH scale runs from one (most acidic)
  to fourteen (most alkaline); the difference of one
  pH unit indicates a ten fold change in hydrogen
  ion activity.

PHYSIOGRAPHY—the elevation, SLOPE, and shape
  of the landform, including its spatial relationships
  with other physical features.

PLAGIOCLASE—common rock-forming minerals in
  the FELDSPAR group,  containing CALCIUM and
  sodium.

PLATINUM COBALT UNIT—measure of the color of
  a water sample defined by a potassium hexach-
  loroplatinate and cobalt chloride standard color
  series.
PLEISTOCENE—the earlier of the two epochs in the
  QUATERNARY period,  relating to the series of
  sediments deposited during this epoch.

PLUTON—a mass of INTRUSIVE  IGNEOUS rock,
  most commonly granite.

PLUTONIC—referring to  rocks,  usually IGNEOUS,
  formed at a great depth.

POPULATION ESTIMATE—a statistical estimate that
  applies to a specific population of TARGET LAKES,
  not only to the sample.

POTENTIOMETER—an instrument  for  measuring
  electromotive forces.  A  source  of  adjustable
  voltage.

PRECAMBRIAN—relating to the  earliest era  of
  geologic  history or the corresponding system of
  rocks, relating to the system of rocks formed before
  the Cambrian Era.
PRECISION—a measure of the capacity of a method
  to provide reproducible measurements  of a
  particular ANALYTE.

PRE-JURASSIC—that period before the  JURASSIC
  period which  was  the  second period  of the
  MESOZOIC Era,  approximately  135-190  million
  years before present.

PRIMARY VARIABLES—variables of foremost con-
  cern  in the Survey (pH, ACID  NEUTRALIZING
  CAPACITY, EXTRACTABLE ALUMINUM, sulfate,
  CALCIUM, DISSOLVED ORGANIC CARBON).

PROBABILITY SAMPLE—a sample in which  each
  unit (lake, in this case) has a known  probability
  of being selected.

PROTOLYTE—that portion of a molecule  that reacts
  with either H+ or OH~ in solution.
                                                                     165

-------
QUALITY ASSURANCE—steps taken to ensure that
  a study is adequately planned and implemented
  to provide data of the  highest quality, and that
  adequate information is provided to determine the
  quality of the DATA BASE resulting from the study.

QUALITY CONTROL—steps taken  during sample
  collection and analysis to ensure that the data
  quality meets the minimum standards established
  by the QUALITY ASSURANCE plan.

QUALITY CONTROL CHECK SAMPLE—a sample of
  known  concentration  used  to verify continued
  calibration of an instrument.

QUANTITATION LIMIT—for each chemical variable,
  a value calculated from BLANK samples  that
  represents the lowest concentration that  can be
  measured with reasonable PRECISION.

QUARTILE—any of the three values that divide  the
  population of a frequency distribution into four
  equal classes, each representing 25 percent; used
  to measure range or variation.

QUATERNARY—the second period of the Cenozoic
  Era  representing  the period from two to three
  million years before present to the present.

QUATERNARY PERIOD—relating to the geologic
  period from the end of the TERTIARY to the present
  time or the corresponding system of rocks.

QUARTZITE—sandstone consisting chiefly of quartz
  grains welded so firmly that,  when broken,  the
  rock breaks  through, rather  than  around,  the
  grains.

QUINTILE—any of the four values (Qi, Q2,  Qs,  Q4)
  that divide  the  population of a frequency distri-
  bution into five equal classes, each representing
  20 percent;  used to provide additional values to
  compare among populations of lakes (see CUMUL-
  ATIVE FREQUENCY DISTRIBUTION).

RAW  DATA SET—the initial data set (DATA SET 1)
  that has received a cursory review to confirm that
  data are  provided in  proper format and  are
  complete and legible.

REAGENT—a  substance added to a water  sample
  (because of  its  chemical reactivity) to determine
  the concentration of a specific ANALYTE.

REAGENT BLANK—a LABORATORY BLANK SAM-
  PLE that contained all  the reagents  required to
  prepare a sample for analysis of silica and total
  aluminum.

REFERENCE VALUE (Xc)—a concentration of interest
  for a given chemical variable.
REGION—for this  report, a  major  area of the
  contermmous  United States where  a substantial
  number of lakes with ALKALINITY less than 400
  //eq L"1 can be found.

RELATIVE BIAS—the expected difference between
  a measured value  and the true value, expressed
  as a percentage of the true value.

REMOTE BASE SITE—location serving as a base of
  operations for sampling crews working more than
  150 miles from the FIELD LABORATORY, requiring
  that samples be flown to the FIELD LABORATORY
  daily.

REPRESENTATIVENESS—a measure of data quality;
  the degree to which sample  data accurately and
  precisely reflect the characteristics of a population.

REQUIRED  DETECTION  LIMIT—for each chemical
  variable, the highest INSTRUMENT DETECTION
  LIMIT allowable in the ANALYTICAL  LABORA-
  TORY contract.

RHYOLITE—light gray volcanic rock with large quartz
  and FELDSPAR crystals in a finer groundmass; the
  fine-grained extrusive equivalent of granite. See
  EXTRUSIVE ROCK.

ROOT MEAN SQUARE—a summary statistic of the
  relative or absolute STANDARD DEVIATION;  a
  pooled STANDARD DEVIATION of the % RELATIVE
  STANDARD DEVIATION (%RSD), calculated by the
  formula: [x2%RSD + SD2%RSDln-1/n']1/2

ROUTINE SAMPLE—the first lake sample collected
  at a site in accordance with standardized protocols.

SANDSTONE—a cemented or otherwise compacted
  detrital  sediment  composed predominantly  of
  quartz grains, the grades of the latter being those
  of sand.

SAS—Statistical Analysis  System, Inc. (Gary,  NC).
  A statistical DATA FILE manipulation package that
  has  data management, statistical, and graphical
  analysis  abilities.  The WLS-I DATA BASE  was
  developed and analyzed primarily using  SAS
  software, and is distributed in SAS format.

SCHIST—a crystalline rock that can be easily split
  into layers;  examples are mica and hornblende,
  which may contain high potassium and f lourine.

SECCHI DISK—a  20  cm-diameter, black-and-white
  disk used to measure water TRANSPARENCY.

SECCHI DISK TRANSPARENCY—a measure of lake
  water clarity determined by  lowering a SECCHI
  DISK into the lake, recording the depth at which
  it disappears from view, then  raising  it until  it
                    166

-------
  comes  into view, and recording the depth; the
  mean value of these depths is then calculated.

SECONDARY  VARIABLES—chemical  variables
  measured  during the survey  considered to be
  important in providing additional data  in quanti-
  fying the chemical status  of lakes, e.g., sodium,
  magnesium, potassium, nitrate, chloride and total
  aluminum.

SEDIMENTARY—referring to  rocks composed of
  particles of other rock transported and deposited
  by water, wind, or ice.

SEEPAGE LAKE—a lake with no permanent surface
  water inlets or outlets.

SHALE—fine-grained  mudstone or claystone with
  laminated  or  layered structure, with  a division
  plane that separates each successive layer.

SHALLOW LAKE—a lake from which a clean sample
  could not be obtained at  1.5 meters below the
  surface but could be obtained at 0.75 meters; these
  lakes were generally less than 3 meters deep.

SLOPE—vertical  distance  divided  by horizontal
  distance.

SOFTWATER—water  with  low concentrations of
  CALCIUM and magnesium salts.

SPARGING—a sample preparation process that
  involved bubbling a gas into an ALIQUOT.

SPECIAL  INTEREST LAKE—in  this survey, a lake
  selected non-randomly which is not part of the
  PROBABILITY  SAMPLE; selection was  based on
  quality and amount  of data available or involve-
  ment in other research programs.

SPLIT SAMPLE—a subsample  of a  field  BATCH
  sample that was sent to a laboratory other than
  an ANALYTICAL  LABORATORY for analysis.
STANDARD  DEVIATION—the square root of the
  variance of a given  statistic, calculated by the
  equation [I x2 - (I x)2]1/2/(n - 1 )1/2.

STANDARD  ERROR—STANDARD  DEVIATION
  divided by the square root of the sample size.

STRATIFICATION FACTORS—factors used to define
  STRATA prior to lake selection; the factors used
  in the WLS-I were  REGION,  SUBREGION,  and
  ALKALINITY MAP CLASS. See STRATUM.

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 between
  the water layers at 1.5 meters below the surface
  and  1.5  meters  above the  lake  bottom. If the
  temperature difference is  also greater than 4°C
  between the water layers at 1.5 meters below the
  surface and 60 percent 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.

SUBPOPULATION—any  defined  subset of the
  TARGET POPULATION.

SUBREGIONS—areas within each REGION that are
  similar in water quality, PHYSIOGRAPHY, vege-
  tation, climate, and soil; used as a STRATIFICA-
  TION FACTOR in the WLS-I design.

SURFACE WATER—streams and lakes.

SURFICAL GEOLOGY—characteristics of the earth's
  surface; especially  consisting of unconsolidated
  residual, ALLUVIAL or glacial deposits lying on the
  BEDROCK.

SYNOPTIC—relating to or displaying conditions as
  they exist simultaneously over a broad area.

SYSTEM DECISION LIMIT—for  each  chemical
  variable, except pH, a value calculated from FIELD
  BLANK data that indicates a concentration (with
  95% confidence) that can be distinguished reliably
  from background levels.

SYSTEMATIC  ERROR—a consistent error introduced
  in the measuring process.  Such error commonly
  results in  biased estimations.

SYSTEMATIC  RANDOM  SAMPLING—a sampling
  technique in which the units in the population are
  ordered. A first sampling unit is randomly  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 protocol.

TERNARY DIAGRAMS—a diagram showing three
  components; for this report, these diagrams were
  used to examine the ratios of major CATIONS and
  ANIONS  (synonymous with trilinear  plots or
  trilinear diagrams).
                                                                       167

-------
TERTIARY PERIOD—relating to the first period of the
  Cenozoic Era or the corresponding system of rocks
  marked by the formation of high mountains.

TITRATION DATA—individual data points from the
  modified GRAN ANALYSIS of ACID NEUTRALIZ-
  ING CAPACITY.

TOPOGRAPHIC MAP—a map showing contours of
  surface elevation.

TRANSPARENCY—the clarity of unfiltered water.

TRAILER BLANK—an ASTM  TYPE  I  REAGENT
  GRADE WATER  sample prepared and processed
  at the FIELD LABORATORY, but analyzed at an
  ANALYTICAL LABORATORY.

TRAILER DUPLICATE—a SPLIT SAMPLE prepared in
  the FIELD LABORATORY.

TRIPLICATE LAKE SAMPLE—the third sample of lake
  water collected  by the  same helicopter crew at
  the same lake immediately after the ROUTINE and
  DUPLICATE SAMPLES in accordance with stan-
  darized protocols; this third sample was used only
  as part of the CALIBRATION STUDY.

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.

TUNDRA—a  level or  undulating treeless plain
  characteristic of arctic  and subarctic regions,
  characterized by a topsoil layer high in  organic
  content above a permanently frozen soil layer.

TURBIDITY—a  measure  of  light scattering  by
  suspended particles in an unfiltered water sample.

TURNOVER—a period of water circulation in lakes,
  occurring when  little or no differences in temper-
  ature are observed with depth.

UPPER CONFIDENCE LIMIT (95%)—a value that, in
  association with a statistic, has  a 95  percent
  chance of  being above  the true value of the
  population parameter of interest.

ULTRAMAFIC—INTRUSIVE  rocks consisting of dark
  coarse grained minerals with relatively low silica
  content.

VALIDATION—process by which data are evaluated
  for quality with reference to the intended data use;
  includes identification of OUTLIERS and evaluation
  of potential SYSTEMATIC ERROR.

VERIFICATION—process of ascertaining the quality
  of the data in  accordance  with  the  minimum
  standards established by the QUALITY ASSUR-
  ANCE plan.

WATERSHED—the  geographic area from which
  surface water drains into a particular lake.

WATERSHED  DISTURBANCE—for this  report,  a
  disturbance  of the  natural environment  in  a
  watershed within 100 meters of the shore as noted
  by field samplers.  Disturbances included roads,
  houses, logging, mining and livestock.

WEAK ACID SYSTEM—a lake in which more than
  10 percent  of  the ANION  charge results  from
  ORGANIC ACIDS or other weak acid ANIONS.

WEATHERING—see MINERAL WEATHERING.

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 single day
  (in this  report, overall within-batch  precision
  includes the effects of sample collection, process-
  ing and analysis; analytical within-batch precision
  includes the effects  of sample  analysis within
  ANALYTICAL LABORATORIES).
                    765

-------
                                           Appendix
Contributed by  W.S. Overton, Department of
Statistics, Oregon  State  University,  Corvallis,
Oregon.

Calibration to Correct for the Effect of
Different Protocols

Use of different protocols involves the risk of bias,
with the  result that data  generated  under  one
protocol may not  be comparable  to data generated
under another. For this reason, the field protocols
of the National Lake  Survey were rigidly standard-
ized, with holding times and  handling  procedures
narrowly prescribed. The need to vary this protocol
in the wilderness areas of the Western Lake Survey -
Phase  I (WLS-I), by ground crew collection,  rather
than the helicopter collection that was the Survey
standard,  caused  some concern about the effect of
the associated changes in protocol, and particularly
in handling and  holding time. Differences in the
protocols for ground crew  and  helicopter  are
discussed in Section 2.5.1.

A strategy was devised between Survey personnel
and the involved Forest Service personnel, to assess
the differences in the two protocols and to minimize
the effect of the change in protocol. This strategy
required a random subset of the wilderness sample
lakes to be "sampled" by both ground crew and by
helicopter, allowing study of the effect of the change
in mode of collection. The strategy followed was to
identify (a) those variables for which  the values
obtained  by the alternate method (ground sample)
could be used interchangeably with values obtained
by the standard helicopter method, (b) those variables
for which the standard value could be predicted
accurately from the alternate value,  and (c)  those
variables for which it was judged that the alternate
method gave values that were too poor to use in
the Survey.

Of the  50 lakes selected for this calibration study,
45 were actually sampled by both ground crew and
helicopter. Duplicate samples  were taken for  these
lakes in both the standard and the alternate  visits.
Visits were scheduled for the same day, but in order
to ensure comparability of the data in the calibration
study, ground  crews were  not apprised of which
lakes were also to be sampled by helicopter. Security
was inadvertently breached in one subregion. The
sample for this subregion was redrawn, and, in order
not to indicate the newly selected lakes, duplicates
were not reassigned for the ground crews. For this
reason, a few sample lakes do not have duplicate
ground sample  values.  Additionally,  some data
values  were removed from the analysis as a result
of data validation.


The  duplicate samples provided valuable evidence
of the repeatability of the measurements, expressed
as measurement error. One sample of each duplicate
pair was sent  to  each  of  the two analytical
laboratories contracted for the WLS-I. Although this
plan caused some difficulty in the interpretation of
measurement error, it created  a welcome opportu-
nity to assess laboratory bias, as reflected in patterns
of differences  in  the results from the two labora-
tories.  Additionally,  analysis  of  audit samples
(Section  4.3)  and field  laboratory  splits provide
information regarding laboratory bias.

The objective of the calibration study was to identify
those variables for which the two protocols generate
interchangeable results,  those for  which the
helicopter values are different,  but predictable from
the  ground crew values, and those for which such
prediction involves unacceptable error, so that the
ground crew data must  be considered unusable in
the population descriptions.
Regression analysis  is  the basic method for
assessment of a linear relationship between the two
values of a variable, representing the same reality,
but determined  by different  protocols. There are
many alternate ways these analyses can be made;
discussion of the rationale for the choices made in
this study are provided by Overton (1986).

As the  ground value is to be used to predict  a
helicopter  value,  the  latter  is  analysed as  the
dependent variable (y) and the ground value as the
independent variable (x). Each lake in the calibration
sample  is represented for each variable by two y's
and most lakes by two x's. Individual y values were
                                               169

-------
regressed on the mean of the x values of that lake.
All regressions were forced through the origin, with
the structural model given by:

                   Y = BX + e

resulting in a simple calibration equation:

                     y = bx

When replacement is implemented, the value of b
is set to 1,  and for calibration by prediction, b is
the least squares estimate of B.

Three  versions of this  model were  used, one
following  each of the following  specifications of
variance of e:
      Model 1. variance proportional to x's
      Model 2. variance inversely proportional to
               x's,
      Model 3. variance homogeneous.
Analysis began with plots of the data, y against x,
in the form of box-duplicate plots (Figures A-1 and
A-2).  Lakes with duplicates in each variable are
represented by a box, one side determined  by the
two duplicate values of y and the other by the two
duplicate values of x. Missing values cause missing
sides  of some  boxes. The  identity  transfer line
indicates the locus on which the two protocols give
identical values.

Inspection of these plots gives a strong visual sense
of the patterns  of the relations.  If the boxes are
arrayed along the transfer line, the relation is close.
One can readily see if the array tends not to pass
through the origin. If the  boxes tend to change size
as a function of x, then one should consider either
Model 1 or 2; if the boxes are homogeneous (in size)
with x, then Model 3 is a candidate. The plots also
provide evidence for erroneous data, and were used
in flagging values for validation. The data ultimately
used  in analysis excluded  those values that were
confirmed to be erroneous.

From  inspection of the graphs, each  variable was
assigned to  one or more  of the above models, and
analyzed for each by least  squares with transfor-
mation appropriate to the model. The model giving
the best  fit was  then   used for that variable
throughout the assessment, with Model 3 prevailing
unless one  of the others  was clearly better. Each
variable was also analyzed with a y-intercept, and
the same error structure;  this allowed assessment
of the consequences of  the choice of the  model
through the origin. Most variables were adequately
analyzed by  Model  3,  with  most  of the rest
conforming  to  Model 1.  One pH  variable (closed
system pH) required Model 2, and  the complement
of this variable (pOH) was analysed by Model 1  to
illustrate that the  assessment  was  identical.  In
general, assessment was robust to model identifi-
cation, and variables that fit one model equally well
as another got essentially the same assessment from
both models.

In establishing criteria for calibration, it is necessary
to consider  the  consequences  of failure of the
alternate protocol value to be identical to the value
from  the standard  protocol.  First,  the alternate
method may be biased, so that  adjustment of the
x values by prediction is required to remove the bias.
This adjustment contributes an additional variance
component to the error of the calibrated value, this
having the effect of additional measurement  error.
Even  if there is  no bias between protocols, it  is
possible  for  the deviation about regression  to  be
substantially greater than the  inherent measure-
ment error,  and  this  also must  be  considered
measurement error of the  alternate  protocol  as
representing the standard. Bias of measurement is
easily detected and  removed by  the methods used
here, but the residual increase in error remains after
correction for bias and is still cause for concern.

This concern is because of the bias in the estimated
distribution, F(x), due to measurement error and to
other  sources of  variation in the  values. The
magnitude of these biases was  investigated  by
Overton and Blick  (1986). They  showed  that
maximum  bias  is  closely  related to the  relative
variance:

                 d2 = 1 + a zm/al.

where a m and a* are measurement variance and
population variance, respectively.

Table A-1 shows the  bias associated with F(x)  as
a function of d2, as determined by Overton and Blick
(1986). They showed  that these values closely
approximate  maximal bias in cases like Model 1,  as
well as Model 3, and  for some  distributions other
than normal.

There is no general standard by which one can select
from  this table a criterion by which to  assess the
calibration;  choice  is  necessarily subjective, and
related  to  the  intended uses of the analyses.  A
Table A-1.   Approximate Bias of F (x) as a  Function of
           Relative Variance, d2. This is Calculated for the
           Simple Case of a Normal Population with Normal
           Measurement Error, and Assessed at ± 1 Stand-
           ard Deviation. (From Overton and Blick 1986)
 d2   1.001  1.01    1.05   1.10   1.20   1.50   2.00
 Bias 0.0001 0.0012 0.0059 0.0115 0.0220 0.0485 0.0811
                        170

-------
     600
                Ca
     300  .
                               300
      8.5
                pOH
      7.0 _
      5.5
      50
           5.5
                              7.0
                 Extractable Al
      20 -
     -10
                                                            650
600
                                                   8.5
                                                            325  -
                                                              8.5
                                                             7.0
                                                              5.5 ,
                                                                 5.5
                                                                         ANC
                     pH
  T
325
                                                                                                          650
                                     I
                                    7.0
                    8.5
          -10
                               20
                                                   50
Figure A-1.    Duplicate box plots of calibration study lakes showing helicopter samples (Y axis) compared to ground samples
              (X axis). Shown are calcium, acid neutralizing capacity, negative logarithm of the hydroxyl ion. pH, and extractable
              aluminum.
                                                                                     171

-------
     740 ,
      70-
                 SO,
                             I
                            70
                                                              6.50
 140
                                                              3.25 -
                                                              O.OO
                            DOC
                                                                   0.00
                                                                                       3.25
                                                                                                          6.50
                DIC
                                                              85.0
                                                              42.5 .
                                                                0.0
                            Conductance
                                                                    0.0
                                                                                       42.5
                                                                                                         850
    14.O
     70
    -0.5
        -05
                NO
                            7.0
14.0
                                                               6.0
                                                               30 _
                                                              -0.5
                           NOl
                                                                  -0.5
                                                                                       3.0
                                                                                                          6.0
Figure A-2.    Duplicate box plots of calibration study lakes showing helicopter samples (Y axis) compared to ground samples
              (X axis). Shown are sulfate, dissolved organic carbon, dissolved inorganic carbon (closed system), conductance,
              and nitrate (regular and expanded scale).
                         172

-------
criterion of d2 = 1.1 is seen associated with maximal
bias of 0.0115 (about 10%), and d2 = 1.01 with bias
of 0.0012 (about 1 % of the value of F(x) at the point
of maximal bias). Relative to the  intended uses of
these data, these levels of bias seem reasonable to
serve as  maximum  acceptable  bias,  and  design
standard bias, respectively.

In the calibration context, prediction error and error
of deviation from  regression simply contribute to
"measurement error," so that assessment of these
errors is appropriately  applied  to  the calibrated
ground values, represented by y. That is, d2 can be
considered a general form of relative error:

      d2 = total variance/population variance

where total  variance  is  the sum  of  population
variance,  measurement error variance, and predic-
tion variance. It is possible,  then, to utilize  several
different values of d2 in the process of choosing the
best treatment of the ground  crew data.  Specifically,
d2 was estimated for the standard protocol, for the
replacement  of y by x, and  for the prediction of y
by the regression of y on x.

From  the regression analyses, the  estimated
variances of prediction were calculated, along with
mean square deviation from regression.  Measure-
ment error variances of y and of x were calculated
from the duplicate values of those quantities, in a
manner appropriate to the model prescribed for each
variable. To complete the computations, the appro-
priate  variance  of x as a replacement for y was
calculated. For Model 3, all of these statistics are
constants over x, but for Models 1 and  2, they vary
with x. The estimated values of d2 are given in Table
A-2; for variables analyzed by Models 1 or 2, these
values are evaluated at the  median  of x in the
calibration sample.

In Table  A-2, most variables can be seen to pose
no  problem due to collection by ground crew. The
first 12 meet the suggested design standard; for four
of those,  prediction  by regression is optimal, but
comparison of the  optimal  and replacement  d2
indicates no appreciable disadvantage from replace-
ment for these four, also.

The second group  of nine variables has somewhat
greater d2, in the range greater than considered
desirable, but still  acceptable. It is of note that for
six  of these, d2 is  nearly as  great for the standard
survey protocol as it is for the optimal calibration.
That is, although the precision of these variables
is nearing the level of concern, this is not due to
the change of protocol, but to inherent  imprecision
in the standard protocol. The exceptions to this are
closed system pH and turbidity; these demonstrate
substantial additional variance due to the calibration
process. Again, for three of the  five for  which
regression is optimal, there is little penalty for using
replacement; closed system pH is the only variable
of the  29  that possibly justifies calibration by
regression.

This leaves  the last group of eight "questionable"
variables, for each  of which  at least  one value of
d2 exceeds the  acceptable value,  some  by a
considerable margin. The first four of these are also
seen not to  owe their undesirable properties to the
alternate  protocol;  in fact, three  of  them appear
"improved" by the regression, two substantially. This
is evidence that the  estimates of measurement error
for these variables are somewhat inflated. The other
four questionable variables were either worsened
by calibration  or have no basis for assessment of
measurement error.

From this analysis, the change in collection protocol
is seen to have created no substantial problem of
comparability of the data. Several variables (turbidity,
closed system  pH,  and  extractable Al)  show an
appreciable  increase in d2 under calibration, and
several (Ca+z, total Al and SiO2) show an inexplicable
reduction in d2. These results can be interpreted in
terms  of  more familiar variance components, as
shown in Table A-3.

In Table  A-3,  it is seen that  several  variables
(turbidity, closed system pH, extractable Al) have very
much greater deviation about regression of y on x
than can be accounted for by measurement  error,
and another group  (Ca+2,  total Al, Si02) have very
much less. These two groups  were identified above
from  Table A-2;  the principal  determinant of
differences  between d? and dl is simply the ratio
between measurement error variance and the  mean
square deviation from regression.

Similarly, the determinant  of the general magnitude
of d2 is the ratio of error variance to population
variance. This follows the definition of d2, but its
significance will  become  more  apparent  in the
following  discussions.  Further, this assessment
relative to the population variation is appropriate
when concern is relative to the impact of errors and
bias  on  population assessment. There  may be
circumstances in  which another criterion is  more
appropriate.

It was earlier indicated  that  measurement  error
estimated from the  calibration sample contained a
component  of laboratory  bias. In the context of
population descriptions, it should contain such a
component,  but not  necessarily in the mix found in
the analysis of duplicates. This  is  clearly  of no
importance for most variables, because the relative
                                                                           173

-------
Table A-2.   Calibration Equations and Decision Criteria for the Twenty-Nine Variables Analyzed in the Calibration Study. The
           Estimated b is Replaced by 1.0 When Replacement is Used; Replacement is Optimal When |t| < 1.
           General Equation: y = bx.

                                                                         d2
Variable3
Within Design Standard
ANC
Ca+2
Mg + 2
DIG, initial
DIG, air-equil.
DIG, closed system
HCO3-
Conductance
F~, total dissolved
S04-2
Na +
K +
t

-0.125
0.575
0.674
0.257
-0.650
0.198
0.833
-0.075
1.277
2.664
-2.682
1.390
b

0.99961
1.00261
1.00293
1.00156
0.99649
1 .00098
1 .00450
0.99967
1.02172
1.01314
0.98849
1.01000
Standard
Protocol

1.00108
1.00156
1 .00063
1.01081
1.00511
1.00131
1.00578
1 .00283
1 .00329
1 .00290
1 .00465
1 .00806
Optimal
Calibration

1.00123
1.00111
1 .00070
1.00818
1.00574
1. 00523°
1 .00500
1 .00447
1.00375
1 .00390
1 .00628
1.01172
Replacement1"









1.00376
1.00395
1 .00646
1.01175
Within Acceptable Standard
Fe
pH, alkalinity
pH, acidity
Depth
Turbidity
pH, closed system
P, total
DOC
cr
Questionable Variables
Al, total
SiO2
pH (air-equil.)
N03-
Al, extractable
True color
Secchi disk
transparency
Temperature
-0.806
-0.304
0.545
0.539
-0.872
-3.674
-1.037
-2.600
1.536

0.437
0.679
-1.783
2.386
0.628
-0.722

-1.702
-2.649
0.98219
0.99960
1 .00090
1.01214
0.98115
0.99313
0.99354
0.97454
1.02240

1.01089
1.01069
0.99619
1.09554
1.05926
0.96388

0.93799
0.93530
1.01696
1.03850
1.06130
d
1.01871
1.00976
1.03894
1.02431
1.02577

1.20802
1.13895
1.15562
1.13662
1.09462
1.41182

d
d
1.02620
1.05526
1.06833
1.05513
1 .07403°
1.06918C
1.03617
1.02909
1 .04009

1.16929
1.08965
1.14249
1.14743
1. 46587°
1. 70104°

1.23717
1.41990





1.07856
1.03621
1.03010
1 .04024



1.14593
1.14779



1.24613
1.47127
aVariables are defined in Volume II, Tables 2-1 and 5-3; units are those used in data analysis (Volume I).
bGiven only when the optimal calibration is predicted by regression.
°These variables apparently have an additional source of variation associated with the change in protocol.
dNot possible to compute.
magnitude of calculated measurement error is so
small, relative to population variance, that substan-
tial  adjustment in  the relative contribution of
laboratory error would have no importance. In only
six variables (all in the last group of Table A-2) is
the  magnitude of  estimated measurement  error
sufficiently large that change of the bias term might
be of influence. Of these, only N0a~ has estimated
laboratory bias of sufficient magnitude (10%, Section
4.5.2) to warrant attention, and even this magnitude
is too small to have an effect, for the levels of N0a~
found in the calibration study.
Three groups  of variables  in Table  A-2  can be
identified as a focus of summary considerations.
Closed system  pH has estimated measurement error
within design standard, but the change in protocol
creates substantially  greater variance  in  this
variable.  Bias  is  slight,  but  bias  correction by
regression is marginally justified. It is indicated that
for this variable,  maximum bias of F(x) would be
changed from approximately 0.007 to 0.006 by use
of calibration (0.99313x) instead of replacement (use
of x). Further, maximum bias would not occur at an
important value of pH; values in the wilderness lakes
are generally well above the  level of concern with
respect to lake acidity.

All the rest of the first two groups (totaling 21
variables) are clearly usable without calibration, and
calibration for closed system pH appears unjustified,
in that little reduction in potential bias is achieved.

The last group of variables in Table A-2  requires
more attention. The  change  in protocol has no
detrimental effect on  NOa", total Al, SiO2, or air-
equilibrated pH and for each of these,  the ground
                         174

-------
Table A-3.   Estimated Variance Components for the Twenty-Nine Variables in the Calibration Study. These Components Provide
           the Basis for the Calibration Methods. All Variance Components are Calculated in Accordance with the Indicated
           Model, and Assessed at the Median of x. The Population Variance is Based Only on the Calibration Sample.
Measurement Error
Variable3
ANC
Ca+2
Al, extractable
pOH
F~, total dissolved
Mg+2
Fe
Al, total
P, total
Secchi disk transparency
pH, closed system
so4-2
DOC
pH, acidity
pH, alkalinity
pH, air-equilibrated
DIG, closed system
DIC, initial
DIG, air-equilibrated
HCO3"
cr
N03~
Na +
K +
Si02
Conductance
Depth
True color
Turbidity
Temperature
y
20.363
21.531
6.4988
0.00222
0.04987
0.9889
17.469
63.150
0.3922
b
0.00222
2.0171
0.01797
0.01051
0.00657
0.02348
0.00334
0.02195
0.01052
82.341
0.4323
0.6007
0.9680
0.2127
0.3192
0.5954
b
9.7222
0.00389
b
X
26.563
24.146
8.2466
0.00728
0.01781
0.8111
22.142
47.144
0.3360
b
0.00726
2.1582
0.01216
0.01792
0.01117
0.01221
0.00228
0.03147
0.02330
81.093
0.2516
0.2259
0.3308
0.1265
0.2055
0.6071
b
5.9211
0.00908
b
Deviation
from
regression
23.331
15.339
31.895
0.01519
0.05650
1.1068
26.863
51.296
0.3604
2.3506
0.01548
2.7079
0.02139
0.01169
0.00933
0.02127
0.01337
0.01660
0.01 180
71.041
0.6760
0.6480
1.3018
0.3086
0.2055
0.9392
5.4354
16.477
0.01536
1.7046
Population
variance
18941.0
13794.6
68.682
0.22696
15.152
1581.35
1030.16
303.57
10.071
10.113
0.22696
695.64
0.73918
0.17151
0.17066
0.15089
2.5576
2.0307
2.0566
14236.2
16.907
4.3973
208.05
26.388
2.2971
210.11
98.790
23.608
0.2079
4.1438
Model
1
1
1
1
1
1
1
1
1
1
2
3
3
3

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3Vanables are defined in Volume II, Tables 2-1  and 5-3, units are those used in data analysis (Volume I); pOH was calculated
 from the closed-system pH value.
bMeasurement error not possible to calculate.
crew data should clearly  be used as  though they
were helicopter  data.  However, this analysis has
disclosed significant relative imprecision in these
four variables, and each  should be examined  for
utility.

The levels of NO3~ are low in the calibration sample,
with a median of 0.45, and greater levels would have
increased the  estimated population variation and
thereby  decreased the value of dz. A moderate    Assessment of SiC>2 is less clear. The median of the
80 percent were below 10.4//gL  1, andthethreshold
of biological concern is approximately 20 /JQ L~1; the
median is below the detection limit (Volume I, Table
4-1). Lastly, the data are accurate enough to indicate
that the level of extractable Al in the wilderness lakes
is very low, even if the precision of the measured
values is questionable. A similar argument applies
to total Al.
change would shift NO3  from the third  group of
variables  to the  second  group, with  acceptable
status.  Further, the levels in  the sample are  low
relative to the detection limit, having the median well
below the assessed value (Volume I, Table 4-1).

Precision of extractable Al is also inherently marginal
in this  analysis, and  is  greatly worsened  by  the
change in protocol, so that one  must question  the
use  of  the  ground data,  even after calibration.
However, the observed values  of extractable Al are
low, relative to levels that would be of concern,  and
near the detection limit. In the calibration sample,
calibration is 2.00 mg L   well above the detection
limit of 0.18 mg L"1 (Volume I, Table 4-1), but close
to the quantitation limit of 2.07 mg L~1  (Volume I,
Table 4-4).  At  the  concentrations of Si02 present
in the wilderness lakes, the precision estimated from
the calibration study (Table A-3) is substantially less
than the precision anticipated from the estimates
from the field  synthetic audit samples  (Volume I,
Table 4-2).

The three physical variables in the set of  suspicious
variables  have somewhat different interpretation.
Secchi  disk transparency  and temperature values
                                                                           775

-------
had no duplicates, so one can only guess that the
protocols  led to substantial  variation in these
variables.  True color did  have the  benefit of
duplicates,  and it  is  seen that protocol  added
substantial variation. True color has coarse resolu-
tion, so one might not expect much more, particularly
with the restricted  range of values found in  this
study. The apparent coarse resolution of temperature
is  surprising; the measurement of temperature is
routine, and its unsatisfactory properties may also
be due only to the narrow  range found in these data.
These three physical variables have low information
content, and so use of the ground data is judged
acceptable.

Reference

Overton,  W.  S., D.  J.  Blick. 1986.  Effects of
  measurement and  other  extraneous  errors  on
  estimated cumulative distributions in the National
  Lake Survey. Presented at the Western Regional
  Meeting of the Institute of Mathematical Statistics,
  July, 1986. Seattle, Washington.
                        176
    . S .  GOVERNMENT PRINTING OFFICES  1987-748-121/67054

-------
O k-
o o
Q_ ^
                               o

                               o
                                         ,-jvirontnental  Protection Agency
                                        . .,',  L" -irary  ( M-L-16)
                                        .  I'^arborn St  eet,  Room 1670
                                  Cm0^.30,. It   60604

-------