xvEPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-92/070
March 1992
User's Guide and
Data Dictionary for
Kenai Lakes
Investigation Project
-------
-------
EPA/600/R-92/070
March 1992
USER'S GUIDE AND DATA DICTIONARY FOR KENAI LAKES
INVESTIGATION PROJECT
March 1992
By
Avis D. Newell
ManTech Environmental Technology, Inc.
U.S. EPA Environmental Research Laboratory
200 SW 35th Street
Con/allis, OR 97333
and
Mark E. Mitch
Department of Environmental Sciences
University of Virginia
Charlottesville, VA 22903
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Con/allis, Oregon 97333
Printed on Recycled Paper
-------
DISCLAIMER
The research described in this report has been funded by the U.S. Environmental Protection
Agency. This document has been prepared at the EPA Environmental Research Laboratory in
Corvallis, Oregon, through Contract No. 68-C8-0006 to ManTech Environmental Technology, Inc.
It has been subjected to the Agency's peer and administrative review and approved for publica-
tion. Mention of trade names or commercial products does not constitute endorsement or recom-
mendation for use.
-------
TABLE OF CONTENTS
Section Page
1. INTRODUCTION 1
1.1 Purpose and Scope l . . . 1
1.2 Objectives 2
1.3 KLIP Target Population 2
2. * DATA COLLECTION 4
2.1 Sample Collection 4
2.2 Processing Laboratory 4
2.3 Analytical Methods 5
2.4 Quality Assurance 6
2.4.1 Bias 6
2.4.2 Precision 8
2.4.3 Detection Limits and Variable Tags 8
2.4.4 Blanks 11
2.4.5 Implications for Data Interpretation 12
2.5 Geographic Variables ..12
3. DATABASE DEVELOPMENT 12
3.1 Transfer Media . 12
2.2 Datasets 13
3.2.1 Dataset 3 (KLIPDS3) 13
3.2.2 Dataset 4 (KLIPDS4) . 13
3.3 Missing Values 14
3.4 Calculated Variables 14
3.5 Flags and Tags 14
4. TARGET POPULATION OF LAKES . 14
4.1 Describing the Target Population 14
4.2 Cumulative Distribution Frequency 15
5. VARIABLE DEFINITIONS . 15
6. ASCII DATASET FORMAT 23
6.1 FORMAT.TXT . 23
6.2 CONTENTS.LIS 25
7. REFERENCES 27
APPENDIX A - KLIP SAMPLE LAKES 29
APPENDIX B - REFERENCE VALUES FOR LRTAP SAMPLES . 32
iii
-------
ILLUSTRATIONS
Figure Page
Figure 1. A map of the Kenai Peninsula showing sampled and target lakes,
and wildlife refuge and wilderness boundaries 3
Figure 2. A cumulative frequency distribution for ANC in the Kenai Lakes
population 16
TABLES
Table Page
-. i • ''
Table 1. Average Percent Bias from Up to Nine Audit Samples Processed in the
Field Laboratory and Sent on to the Analytical Laboratory 7
Table 2. Average Percent Relative Standard Deviation (RSD) of Duplicate" Samples -
for the Major Analytes of the Kenai Lakes Investigation Project 9
Table 3. Analytical Detection Limits for Appropriate Variables 10
Table 4 Results of Analyses on Field and Filtered Blanks for Those Variables with
Blank Results Greater than the Detection Limit for the Corresponding
Analyte :..'.... 11
IV
-------
1. INTRODUCTION
1.1 PURPOSE AND SCOPE
In 1984, the U.S. Environmental Protection Agency (EPA) implemented the National Surface
Water Survey (NSWS) (Linthurst et al., 1986; Landers et al., 1987; Kaufmann et al., 1988) as part
of the Aquatic Effects Research Program (AERP). The AERP conducted several integrated studies
in areas containing surface waters considered potentially sensitive to change as a result of acidic
deposition. The AERP addressed five major environmental policy issues:
1. The present status and extent of acidic and low alkalinity surface waters in the United
States
2. The extent and magnitude of past change < ,
3. The change to be expected in the future under various rates of acidic deposition
4. The maximum rates of deposition below which further change would not be expected
5. , The rate of change or recovery of aquatic ecosystems if deposition rates decreased.
The NSWS focused its assessment on lakes and streams located in the contiguous United
States. Since the majority of the systems examined in the NSWS receive moderate to high levels
of acidic deposition, it is difficult to evaluate the role of natural factors in controlling the chemistry
of aquatic ecosystems. Therefore, the EPA implemented a project to collect data on lakes in the
Kenai Peninsula of Alaska, an area expected to receive low levels of acidic deposition.
The data contained on the accompanying tape or floppy diskettes were- collected during the
Kenai Lakes Investigation Project (KLIP) in the late summer of 1988 (Eilers et al., in press). Like
other components of the NSWS, KLIP is based on a probability sample from an explicitly defined
population of surface waters. Data were collected during fail overturn, considered to be a repre-
sentative "index" period. Sample information was then extrapolated to represent the target lake
population in Alaska's Kenai Peninsula.
This database guide provides a brief overview of the survey and the KLIP database.
Detailed information on KLIP results is found in Eilers et al. (in press). This document also
summarizes the sampling (Section 2.1) and analytical methods (Section 2.3), sources of geo-
graphic information (Section 2.5), and precision and accuracy results from quality assurance (QA)
analysis (Section 2.4). The datasets are described in Section 3 and their formats in Section 6.
The variables are defined in Section 5, and Appendix A contains a list of the lakes and their
chemistry. Appendix B provides reference values taken from the Long Range, Transport of Air-.
borne Pollutants (LRTAP) Project audit samples. ,
1
-------
The results of this study showed that over 90% of the target population lakes are seepage
systems, having no surficial inlets or outlets and only a small direct terrestrial drainage area. The
median lake surface area was 8.3 ha and the median watershed area was 79.3 ha. Watershed
processes probably have a minimal influence on these systems, except in cases of extreme dis-
turbance. In turn, these systems are expected to be highly responsive to groundwater inputs.
When such inputs are small, these systems are likely to be responsive to atmospheric inputs of
acidic materials due to lack of base cations from the weathering of soils and tills.
1.2 OBJECTIVES
From a regional perspective, little is known about the chemical character of Alaskan lakes,
and the KLIP survey provided an opportunity to collect such information in one part of Alaska.
We selected the Kenai Peninsula lakes in part because the surficial morphology of the region is
similar to that of the midwest, yet the Kenai receives considerably less acidic dejDosition. The
area contains many seepage lakes in the relatively flat topography characteristic of glacial till
deposits. We hoped that these lakes would provide a good comparison to the morphologically
similar but moderately impacted Upper Midwest lakes. In addition, we were curious about the
possible impacts from the Nikiski Industrial complex, located on the northwest corner of the area
studied, and thus have included the distance from each lake to this area. Results from this
survey are available in Eilers et al. (in press).
1.3 KUP TARGET POPULATION
KLIP focused on a population of 902 lakes portrayed on 1:250,000-scale U.S. Geological
Survey (USGS) topographic maps of the Kenai Peninsula in an area delineated by the eastern
and southern extent of glacial deposits on the lowland area of the peninsula (Figure 1). The
sampling frame was generated by assigning each lake a unique identification number. Lakes
were numbered consecutively, starting in the northwest corner of the peninsula and working east.
This procedure was repeated from north to south until all lakes were numbered.
A subsample of over 60 lakes was then systematically selected for field visitation, using a
random start. Once lakes were identified for sampling, detailed geographic data were recorded
for each selected lake (watershed area, lake area, proximity to roads, etc.) and used to determine
access for sampling. During pre-sampling field reconnaissance, four of the selected lakes were
determined to be dry, and one lake had nesting swans. During sampling, two sites were found
not to be lakes, and one other had nesting swans. We agreed not to sample Jakes with nesting
-------
o
uu
O _
(0
0)
^
(0
•u
c
o
(0
(0
(U
^
0)
c
(0
0)
D)
DC
0)
T3
(0
of
0)
^
>
(0
Q.
E
TO
(0
(0
a)
a.
'n
I
o
o
a.
0)
3
D)
u.
-------
swans, following U.S. Fish and Wildlife management policy regarding disturbance of the nest
sites. Thus, 59 lakes were successfully sampled, representing a target population of 819 lakes
(59 lakes x 13.8769, the WEIGHT for each lake). The target population of lakes characterized by
KLIP is considered to be only those represented on 1:250,000-scale topographic maps that do
not have nesting swans. KLIP data should not be used to infer conditions of lakes either not
depicted on these maps or outside of the sampling area.
Unlike the other regions studied by the NSWS (Linthurst et al., 1986; Landers et al., 1987), a
stratified sampling design was not used in KLIP. Few existing lake chemistry data were available
for the delineation of strata on the Kenai Peninsula, so the area was sampled as a single stratum.
This procedure has simplified data summary calculations, as each observation in KLIP represents
13.88 lakes in the target population, as indicated by the weighting factor (the variable WEIGHT in
the datset).
2. DATA COLLECTION
2.1 SAMPLE COLLECTION
Field crews sampled 59 lakes during August of 1988. Lakes were accessed via float plane,
helicopter, or canoe. Secchi disk measurements were made, followed by a water temperature
profile to determine if the lake was thermally stratified. A Van Dorn™ sampler was used to collect
a 6-L water sample at a depth of 1.5 m at the deepest point in the lake. From this, a 4.5-L Cubi-
tainer™ was filled and a 60-cc polypropylene syringe sample was withdrawn. Both were stored
on ice until arrival at the processing laboratory. Duplicates, collected as part of the quality
assurance (QA) program, consisted of field splits; two Cubitainers™ and two pH syringes were
filled from a single Van Dorn™ sample. National Surface Water Survey field methods were
followed (Chaloud et al., 1987).
2.2 PROCESSING LABORATORY
A processing laboratory was set up in the U.S. Fish and Wildlife facility to stabilize water
samples before shipment to the analytical laboratory. AH samples were processed within 6-12
hours of collection. The samples were processed into 0.45-jum filtered aliquots for cation, anion,
and dissolved organic carbon (DOC) analysis. An unfiltered aliquot was prepared for acid neutral-
izing capacity (ANC) and laboratory conductance measurements. Cation and DOC aliquots were
preserved with ultra-pure HNO3 and H2SO4, respectively. All of these aliquots were shipped to
-------
the EPA Region X laboratory in Manchester, Washington, via overnight courier, where they were
later analyzed. Samples were stored on ice, or refrigerated at 4°C at all times, including shipping,
with the exception of sample processing.
Measurements of specific conductance, color, and closed-cell pH were made at the proces-
sing laboratory, using methods developed as part of the NSWS (U.S. EPA, 1987). Color was
determined on filtered samples using a visual comparator. Conductance was measured on an
unfiltered aliquot. Syringes for closed-cell pH samples were filled directly from the Van Dorn™
sampler in the field; pH was measured in the processing laboratory, on samples equilibrated to
room temperature. Replicate values were averaged to obtain PHSTVL, the variable for field pH in
this database.
2.3 ANALYTICAL METHODS
Inductively coupled plasma (ICP) spectroscopy was used to analyze for Ca, Mg, Na, Al, Fe,
Mn, and Si. Potassium was analyzed using atomic absorption spectroscopy. Sulfate, NO3", and
CI" were analyzed using ion chromatography. DOC was determined by infra-red (IR) spectros-
copy after persulfate oxidation on a Dohrman™ carbon analyzer. Dissolved P was determined on
an autoanalyzer using ammonia-persulfate digestion followed by single-reagent ascorbic acid
colorimetry (U.S. EPA, 1979; Method 365.2). Dissolved ammonia-nitrogen was determined by an
automated colorimetric phenate method (U.S. EPA, 1979; Method 350.1).
ANC was determined by a double eridpoint titration. Although this method is generally not
considered as accurate as the Gran titration method that was used in other NSWS studies, the
results of three audit samples ranging from 50 to 814^eq L"1 ANC were comparable to Gran
titration reference values for these audit samples.
pH was measured in the analytical laboratory. The field pH value, PHSTVL, results in better
ion balances and calculated conductance values. For KLIP, the PHSTVL measurements were
made within 12 hours of collection, whereas the laboratory pH values (PH_LAB) were measured
after shipping; holding times on these ranged up to three months.
Both laboratory (U.S. EPA Manchester Laboratory) and field conductance (measured in the
field laboratory) values are available. Based on the results of 9 audit samples, 10 duplicate lake
samples, and calculated versus measured conductance for the lake samples, the laboratory con-
ductance (CONDLAB) values appear to be both more accurate and more precise than field con-
ductance (CONDFLD) measurements.
-------
2.4 QUALITY ASSURANCE
Three types of QA samples were used during KLIP. These included 9 audit samples for
assessing bias, obtained from the Long Range Transport of Airborne Pollutants '(UTTAP)
Interlaboratory Study in Canada (Aspila, 1989), duplicate samples from 11 of the 59 lakes
sampled for assessing precision, and 12 deionized water blanks to identify possible occurrences
of contamination. Of the 12 blank samples, 8 Cubitainers™ were filled in the field and treated as
samples upon return to the laboratory; these were considered field blanks. The remaining four
were filtered in the same way as lake samples and field blank samples, but had originated in the
processing laboratory. Results for these three types of QA samples are presented in the tables
included with the discussion in the following sections.
The following tables present average percent bias and precision (as relative standard
deviation, which is the same as the coefficient of variation). These tables are intended only as a
broad summary of data quality. As indicated in Section 3.2, the results for all of the QA samples
are included in Dataset 3, and the reference values for the audit samples are included in
Appendix B, so that more detailed analysis of the QA data by any user is possible.
2.4.1 Bias
Nine audit samples were processed and analyzed during KLIP. The samples were provided
by the LRTAP Interlaboratory Study in Canada (Aspila, 1989, study number 16). These LRTAP
samples, consisting of natural lake, stream, or rain water, had been sent to over 50 laboratories
for analysis. The median values of this large number of analyses were then used as reference
values to estimate percent bias. Quality assurance objectives (QAOs) were defined before the
project was begun. For bias, these objectives were set to ± 10% for all variables, with the
exception of conductance, which was set at ± 5%. The average percent bias is shown in
Table 1. Not all of the LRTAP sample concentrations corresponded to the concentration ranges
of the KLIP lake samples; only the audit samples that bounded the KLIP lake concentration range
are included in Table 1.
The worst performance was for K, for which 5 of 9 audit samples were slightly high. The K
concentrations of both the audit samples (< as^eq L"1) and KLIP lakes (< 37 ^eq L"1) were quite
low, however, and the absolute differences between audit values and laboratory results for all
samples exceeding the QAO were < 3 ^eq L"1. This performance is within the analytical
precision for K, and is thus an acceptable result, so the QAO is overly stringent.
-------
Table 1. Average Percent Bias from Up to Nine Audit Samples Processed in the Field
Laboratory and Sent on to the Analytical Laboratory3
Variable
ANC
pH
Conductance
Ca
Mg
K
Na
S042'
cr
N03-
NH3
Color PCU
DOC
n
3
4
7
9
9
9
8
9
8
6
9
5
9
Average
% Bias
6.5%
-2.6%
1.85%
2.8%
4.5%
18.4%
-3.3%
-4.2%
2.0%
6.7%
-32%
138%
6%
Minimum and
Maximum
% Bias
-17.2 0.8%
-5.2 .2%
-0.5—8.5%
0—6.7%
0—6.7%
-2.1—40.6%
-12.9—5.9%
-14.5—4.7%
-2.8—8%
-44—82%
-72—3.4%
20—400%
-22—35%
Concentration
Range of Audit
Samples
50— 814^eq L'1
5.56—7.75 pH
units
6.0 — 94 fiS cm"1
15— 2138,ueq L"1
4—794 ^eq L'1
3.5—23 fieq L"1
13— 264 fieq L'\
8—782 ,ueq L'1
6—306 peq L'1
1— 15 ^eq L"1
0.4 — 25 ,Meq L'1
1—110 PCU
0.1—11 mgL'1
Concentration
Range of Audit
Samples
20— 1738/*eq L'1
5.53—8.88 pH
units
8.3— 175^5 cm'1
11— 1367 ^'eq L'1
17.5— 389 [teq L'1
2—37 ,«eq L'1
24 — 219,ueq L"1
1b— 46^eqL-1
23— 189/teq L'1
0.4b— 3.4 ,aeq L'1
0.4b— 6.2 ^eq L'1
5—100 PCU
2.3—17.5 mg L'1
Also shown on the table are the minimum and maximum percent bias, the concentration range of the audit samples,
and the concentration range of the Kenai Lake samples.
Minimum concentration is the analytical detection limit.
The %bias values that exceed the quality assurance objective (>10%) for these samples represent only a 1 ,ueq L"1
absolute difference between the audit result and the reference value.
-------
In addition to the calculation of percent bias to determine if bias was present for any of the
variables in the KLIP dataset, the Wilcoxin signed rank test (Hogg and Tanis, 1983) was applied to
see if there was a consistent deviation from the median value for any the variables. These results
showed no consistent bias for any variables.
2.4.2 Precision
Of the 59 lakes sampled, 10 duplicate samples were taken, with 1 or 2 taken on each day of
sampling. Precision values are determined by calculating the relative percent standard deviation
(%RSD), which is the same as the coefficient of variation, and is equal to the standard deviation
divided by the sample mean, and multiplied by 100. These results are presented in Table 2. The
QAOs for precision were set at 5% RSD for most variables, with the exception of ANC (10% RSD),
NO3" (10% RSD), and DOC at concentrations < 5 mg L"1 (10% RSD). The precision criterion for
color was an absolute difference of 5 PCU, and was not based on relative standard deviation.
The three ANC duplicate pairs that exceeded the QAO were at low ANC concentrations, and
reflected acceptable absolute differences of 10—15 ^eq L'1 (26.5% RSD at 31 ^weq L'1 ANC;
15.7% RSD at 36 fieq L"1; and 13% RSD at 44 ,«eq L"1). The two pairs of Ca duplicates that
exceeded the QAO were also at low concentration and reflected < 5 ^ueq L"1 absolute difference
Within pairs (10.8% RSD at 20/ieq L'1 Ca, and 12.5% RSD at 17/^eq L'1 Ca). Thus, although
some duplicate pairs for ANC and Ca exceeded their QAO, the QAO was inaccurate in defining
an acceptable quality for these variables at these levels.
The two pairs of NH3 and Mn exceeding the precision goals were at very low concentrations
(< 1 fieq L"1 NH3 and < 10 ^wg L"1 Mn). Two sets of duplicate values for Fe had poor precision
(47% RSD at 10 ^ug L"1, and 20% RSD at 50 fig L"1). Three duplicate pairs of color were outside
the acceptable limits; however, color values were generally very low (< 100 PCU); 75% of the
color values were < 30 PCU, the definition of low color in the NSWS. Color was the vari-
able of least interest in the KLIP Project.
NO3~ precision was poor, but difficult to quantify, as all three pairs with detectable NO3" had
high %RSD values. Three of the DOC pairs differed by as much as 2.5 mg L"1, and exceeded the
precision goal of 10% RSD.
2.4.3 Detection Limits and Variable Tags
The analytical laboratory determined the analytical detection limits (Table 3), and did not
report values for results that were below these levels. Observations with low concentrations were
8
-------
Table 2. Average Percent Relative Standard Deviation (BSD) of Duplicate Samples for the
Major Analytes of the Kenai Lakes Investigation Project3
Variable
ANC
PH
n
10
10
Average
% Bias
6.5%
0.6%
Minimum and
Maximum
% Bias
0—26.5%
0—3.4%
Concentration
Range of Audit
Samples
20— USS^eq L'1
5.53—7.96 pH
units
Concentration
Range of Audit
Samples
20— 1738,weq L"1
5.53—8.88 pH
units
Conductance 10
0.2%
0—1.0% 9.5—126 fiS cm"1 8.3—175 fiS cm
-1
Ca
Mg
K
Na
S042'
cr
N03-
NH3
Color PCU
DOC
Al
Fe
Mn
Si
10
10
10
10
10
10
10
10
10
10
10
10
10
10
3.4%
0.2%
2.5%
1.3%
2.8%
1.4%
30%
8.5%
12.4%
7%
1%
9.4%
11.3%
0.3%
0—12.5%
0—0.7%
0.1—8.7%
0.5—3.3%
0—8.3%
0^.3%
0—133%
0—33%
0—47%
0.4—22%
0—6.4%
0-^7.1%
0—78.5%
0—2.7%
17— 895 ^aeq L'1
21— 305/ieq L'1
8.4-^34 /ieq L'1
25— 219,aeq L'1
1 .0—46 jueq L'1
22— 189/ieq L'1
0.4 — 6 fieq L'1
0.4—6 ^ueq L'1
5—125 PCU
3.7— 17.5 mgL'1
25—231 fig L'1
7—521 fig L'1
2—37 .ugL'1
40—2395 fig L'1
11— 1367/teq L"1
17.5— 389 ,ueq L'1
2—37 /
-------
Table 3. Analytical Detection Limits for Appropriate Variables
Variable
Analytical Detection Limit
Ca (weq L"1)
Mg («eq L"1)
Na (ueq L'1)
K (weq L"1)
S042' (ueq L'1)
OP («eq L'1
N03- («eq L-1)
NH3 (aeq L"1)
P C«g L-1)
DOC (mg L'1)
0.05
0.08
0.9
1.5
1.0
2.8
0.4
0.4
7.0
0.1
Si («g L'1)
Al («g L'1)
Fe (fig L"1)
Mn («g L"1)
40
25
7
2
10
-------
reported as the detection limit value and tagged; thus, no zero or negative values occur in the
dataset. Tag values assigned by the analytical laboratory for each chemical variable appear in a
separate variable, and are used to qualify the values for the chemical variables. Two tag values
were employed, 'U' and 'M'. The 'U' value indicates that the sample was analyzed, but no analyte
was detected in the sample. A tag value of 'M1 indicates that the analyte was detected, but not
quantified; in other words, the analyte occurred at a trace amount.
2.4.4 Blanks
Two kinds of blanks were included in the KLIP project; deionized water was taken into the
field and used to fill a Cubitainer™ each day. These were not put through the Van Dorn™ water
sampler, as there was no means of carrying enough water into the field to rinse the sampler prior
to taking a blank sample. Blanks were also filtered in the laboratory, without having been
processed in the field. The results of all blank samples for Mg, Na, Si, Al, Mn, Fe, SO42', NO3",
and CI" were at the detection limit values for these variables. Results for the remaining variables
are shown in Table 4. None of the results for these blanks were high enough to be of concern.
The blank results for Ca, K, NH3, and DOC were very close to the detection limit values, and far
below values for the KLIP lakes.
Table 4. Results of Analyses on Field and Filtered Blanks for Those Variables with Blank
Results Greater than the Detection Limit for the Corresponding Analyte
ANC(weq L'1)
Conductance («S cm"1)
Ca ("eq L'1)
K («eq L'1)
NH3 (aeq L'1)
DOC (mg L'1)
Mean
5.99
1.08
0.34
1.43
0.37
0.14
Standard
Deviation
6.38
0.09
0.35
0.29
0.06
0.04
Number of
Blanks,
11
11
11
11
12
12
Detection
Limit
n.a.a
n.a.a
0.05
0.77
0.36
0.1
No detection limits are available for these variables.
11
-------
2.4.5 Implications for Data Interpretation
Evaluation of the QA data from the KLIP Project indicates that only the duplicate values for
NO3" were of questionable quality. Values for NO3" were below detection limit for so many of the
KLIP lake samples that it is difficult to make many inferences concerning NO3", thus the poor
precision is of little concern. Percent bias for NO3" was within acceptable limits; so although the
data may not be precise, they are not biased. Color data for KLIP were not good, but are gener-
ally of little interest due to the subjective nature of obtaining color data.
As demonstrated in this text, the KLIP data are of very high quality. Quality assurance
samples that failed to meet their QAOs were generally of an acceptable level of quality to meet
the study objectives. Inaccurate selection of QAOs at variable concentration levels resulted in the
identification of some precision and audit data that did not meet the QA objectives, but were in
fact at acceptable performance levels.
2.5 GEOGRAPHIC VARIABLES
Several geographic variables occur in the database and were obtained from 1:25,000-scale
topographic maps of the region: latitude, longitude, watershed area, lake area, lake name, lake
elevation, and lake hydrologic type [seepage (no inlets or outlets) or drainage (outlet present)].
Measurements of lake area, watershed area, distance to the coast, and distance to the refineries
at Nikiski were determined by electronic planimetry.
3. DATABASE DEVELOPMENT
3.1 TRANSFER MEDIA
The database exists in two formats: as SAS transport datasets (SAS, 1988; version 6.04),
and as ASCII files on 51/4-inch, low-density (360 Kb) floppy diskettes. The floppies contain (1) two
datasets: datasets 3 and 4, KLIPDS3 and KLIPDS4, described in Section 3.2, and (2) the data-
base descriptions in two files: CONTENTS.LIS, a listing of the variable names and definitions,
and FORMAT.TXT, a description of the ASCII file columnar format. These are further described in
Section 6. The SAS transport datasets have been created using PROC CPORT, and can be
transferred back into SAS datasets using PROC CIMPORT (SAS, 1988).
12
-------
3.2 DATASETS
One of the two KLIP datasets, KLIPDS3, includes all data collected during the KLIP project:
information on the 62 lakes that were visited out of the 67 lakes selected for sampling, and all QA
samples, such as audits and duplicates. The five lakes determined to be unsuitable for sampling
during pre-sampling reconnaissance are not included. The other dataset, KLIPDS4, includes
information only for the 59 sampled lakes used to characterize the KLIP target population.
Dataset 4 is the easiest to use for characterizing the Kenai Lakes or for analyzing inter-variable
relationships. Dataset 3 documents the results of all analyses performed for KILP, and may be
used to characterize the data quality. Further differences between the two datasets are described
in the following sections.
3.2.1 Dataset 3 (KLIPDS3)
Dataset 3 (KLIPDS3) contains verified data and includes QA data in the form of audit results
and analysis of the laboratory (filter) and field blanks and the duplicate pairs. Verified data
indicate that the numbers from the field forms and analytical laboratory were properly transferred
to the database. All other possible errors, such as analytical errors or recording errors, are
included in Dataset 3. Data found to be in error through validation are indicated as such in the
comment field of KLIPDS3, but the original values remain. Validation was performed by exam-
ining ion balances, comparing calculated to measured conductance, and examining relationships
between variables for outliers in much the same way as described in U.S. EPA (1992).
Field blanks are identified as such only in the sample code (SAMPCODE) variable. These
samples have LAKEJD and LAKENAME values that indicate the site at which the field blanks were
taken. Filter blank samples are indicated as 'BLANK' in LAKENAME, and as 'BFIL1 in
SAMPCODE, Two SPLIT samples are included in KLIPDS3. These were part of a comparison for
another study, and are not indicative of KLIP performance.
3.2.2 Dataset 4 (KLIPDS4)
Dataset 4 (KLIPDS4) contains only data from the lakes successfully sampled. This is the
dataset used for regional characterization of the Kenai Lakes. Duplicate values have been aver-
aged, and extreme outliers, identified during the validation process (Section 3.2.1), have been
omitted or replaced. Of the three values found to be in error, two were from duplicate samples.
Only the acceptable duplicate was included in KLIPDS4, not an average of the two duplicate
13
-------
values. The third, an ANC value, was estimated from the relationship between ANC and the sum
of nonmarine base cations minus nonmarine suifate. The error about this relationship was used
to estimate a random error component for the datum and incorporated into the estimate (Little
and Rubin, 1987).
3.3 MISSING VALUES
Missing values are represented by the number -999.000 for the card-image ASCII format.
Standard SAS notation for missing values, a"." for numeric values and blank space for character
values, is used in all SAS files. The only missing values in KLIPDS4 occur in the variables
WSHED (watershed area), LAKESIZ (lake area), and WTEMPB (water temperature 1.5 m above
the lake bottom), and for the various TAG variables.
3.4 CALCULATED VARIABLES
A few variables were calculated as part of the validation process and remain in the final
datasets. These are hydrogen and hydroxyl ion concentrations, the sum of anions, the sum of
cations, the anion deficit, the calculated conductance, and the organic ion present as calculated
from the Oliver model (Oliver et al., 1983). The equations used to calculate these variables are
summarized in Section 5.
3.5 FLAGS AND TAGS
Qualifiers from the analytical laboratory are indicated in the tag variables for each analyte.
Values that were determined to be in error through charge balances or other relationships among
variables are identified in the comment field. Original values are included in Dataset 3, and values
that are representative of the Kenai Lowland lake population (described in Section 3.2.2) are
included in Dataset 4. The tag values are defined in Section 2.4.3.
4. TARGET POPULATION OF LAKES
4.1 DESCRIBING THE TARGET POPULATION
The KLIP database structure facilitates examination of the target population in a variety of
ways. Estimates of subpopulations of lakes can readily be calculated and the distribution of the
14
-------
entire target population of systems can be examined. For example, if one is interested in exam-
ining information only for seepage lakes, the data can be subset using the appropriate variable (in
this case, LAKE_TYP='S'). The number of lakes described by each sampled lake is calculated by
multiplying by the weighting factor (WEIGHT). Since KLIP was not based on a stratified sampling
design, each lake is equally weighted (e.g., the variable WEIGHT is the same for each lake).
4.2 CUMULATIVE DISTRIBUTION FREQUENCY
Population distributions can be generated for measured attributes using the general estima-
tors (e.g., number of lakes, watershed area, elevation) and are presented using the cumulative
distribution function, F(x), sometimes called a cumulative proportion. The distribution, F(x),
shown in Figure 2, is interpreted as the proportion of target lakes in the population having the
attribute X < x. To read this figure, select a value of x of the attribute X, along the hori-
zontal axis (in this example, ANC) and read the Y-axis value, F(x), the cumulative proportion of
lakes at this value. In this example, to determine the proportion of lakes with ANC < 200
fieq L"1, locate the ANC value of 200 along the X-axis, then read the corresponding F(x) as 0.7.
The estimated number of lakes (N) with ANC < 200 /*eq L"1 can be calculated by multi-
plying F by Ntota| (the total number of lakes in the population of target lakes). For our example,
this is (0.7) (819) = 573 estimated lakes with ANC < 200 ^eq L'1. We can also choose a
value of F(x), such as the median at F(x) = 0.5, and read off the estimated x value (in this case,
93 ,«eq L"1). Thus we can estimate the median (or other percentile) ANC value.
Distributions are sometimes presented as an inverse or descending cumulative proportion,
1-F(x). For these, the distribution is read as the estimated proportion of lakes having values > x
(X > x). For the lake surveys, the inverse cumulative proportions have been used to describe
distributions where there is more interest in the maximum values than the minimum values. For
example, because SO42" can be indicative of acidic inputs, we were more interested in what
proportion of,the population had SO42" above a certain value. Similarly, higher elevation lakes
were expected to have greater inputs of atmospheric acids, so we were more interested in esti-
mating the proportion of lakes found at greater than a given elevation, rather than the proportion
with elevation below a given value.
5. VARIABLE DEFINITIONS
Following is an alphabetical list of all the variables in the KLIP dataset, and their definitions,
including units used for measure, and a brief description of how the measure was made. The
15
-------
16
-------
TAG variables indicate whether the variable value is accompanied by 'U' or 'M', which are defined
in Section 2.4.3.
AL
ALK_TOT
AL_T
ANDEF
-1
ANSUM
CA
CA_T
CATSUM
CL
CL_T
COLVAL
COMMENT
-1
1-1
-1
-1
-1
-1
PCU
Total dissolved aluminum measured using inductively coupled
plasma, with a detection limit of 25 pg L"1.
The ANC of the water measured by double endpoint titration.
Analytical laboratory assigned tag for Al (U or M).
Anion deficit calculated as cations minus anions, using the equation:
ANDEF = CATSUM - ANSUM.
The sum of anions calculated as follows:
ANSUM = SO4 + NO3 + CL + SCO3 + OHMINUS,
where OHMINUS = (10pH"14)(106).
Dissolved calcium measured by inductively coupled plasma. The
detection limit was 0.05 «eq L"1.
Analytical laboratory assigned tag for Ca (U or M).
The sum of cations calculated as follows:
CATSUM == CA + MG + Na + K + NH4ION + HPLUS,
where NH4ION = NH3(10'pH)/[5.012 x 10'10 + (10'pH)].
Dissolved chloride concentration measured by ion chromatography.
The detection limit was 2.8 ^weq L"1.
Analytical laboratory assigned tag for Cl (U or M).
The true color measured with a visual comparator, in platinum-cobalt
units (PCU).
A 40-character string for notations concerning each observation, this
variable has values only in Dataset 3 (KLJPDS3).
17
-------
CONDCAL /*M cm
-1
CONDFLD ftM cm
-1
CONDLAB fiM cm
-1
Conductance calculated from the measures of the major ions and
their respective contributions to conductance as follows:
CONDCAL = [(CA x 59.47) + (MG x 53.0) + (NA X 50.08) + (K X
73.48) + (NH4ION x 73.5) + (HPLUS x 349.65) + (SO4 x 80.0) +
(SCO3 x 44.5) + (Cl x 76.31) + (NO3 x 71.42) + (OHMINUS x
198)]/1000,
where NH4ION = NH3(10'pH)/[5.012 x 10'10 + (10'pH)]
OHMINUS = (10pH'14)(106).
Specific conductance measured in the field laboratory using a Fisher
portable conductance meter, on samples that had equilibrated to
room temperature. QA data suggest that CONDLAB was a more
accurate measure of specific conductance.
Specific conductance measured in the analytical laboratory, also
measured at room temperature. QA data suggest that this measure
was more accurate than the field conductance value.
DATE
DIS REF
DOC
ELEV
FE
FET
DDMMMYY The date on which the sample was taken from the lake, in the format
day, month, year (i.e., 18AUG88).
km
mgCL
-1
DOC_T
DS COAST km
m
1-1
The distance from the lake to the closest refinery, measured from
1:25,000-scale USGS topographic maps.
Dissolved organic carbon, measured by IR spectroscopy after per-
sulfate oxidation on a Dohrman carbon analyzer. Detection limit was
0.1 mg L'1.
Analytical laboratory assigned tag for DOC.
The shortest distance from the lake to the coast, measured from
1:25,000-scale USGS topographic maps.
The lake elevation above mean sea level, as determined from
1:25,000-scale USGS topographic maps.
Dissolved iron, measured by inductively coupled plasma. The
detection limit was 7 pg L~1.
Analytical laboratory assigned tag for Fe (U or M).
18
-------
HPLUS
-1
K
K_T
LAKENAME
LAKE ACC
LAKE ID
-1
LAKE SIZ ha
LAKE TYP
LAT DMS DMS
LONG DMS DMS
MG
fieq L
-1
Hydrogen ion concentration determined from PHSTVL as follows:
HPLUS = (10-pH)(106).
Dissolved potassium measured by atomic absorption spectroscopy.
Detection limit was 1.5 fig L"1.
Analytical laboratory assigned tag for K (U or M).
The lake name taken from 1:25,000-scale USGS topographic maps.
A code indicating how the lake was accessed, where:
HE = Helicopter
FW = Fixed wing aircraft
CA = Canoe.
A seven-digit code used to identify the lake. All KLIP lakes start with
7AO-'. This convention is from the NSWS, where the region is repre-
sented by the first number, the subregion by the letter, and the
stratum in the third column. Here, Alaska is Region 7, the Kenai
Peninsula is subregion A, and, as the sample was not stratified, the
stratum is always 0.
The lake area in hectares as measured from 1:25,000-scale USGS
topographic maps.
A variable indicating lake type, as estimated from 1:25,000-scale
USGS topographic maps. There were 10 drainage lakes in the
survey.
D = Drainage lakes, with inlets and outlets present.
S = Seepage lakes, where inlets and outlets were absent.
r
The latitude of the lake in degrees, minutes, and seconds,
determined from 1:25,000-scale USGS topographic maps.
The longitude of the lake in degrees, minutes, and seconds,
determined from 1:25,000-scale USGS topographic maps.
Dissolved magnesium concentration measured by inductively
coupled plasma. The detection limit was 0.08 /ieq L"1.
19
-------
MG_T
MN
MN_T
NA
NA_T
NH3
NH3JT
N03
NO3_T
ORGION
PHSTVL
1-1
-1
-1
-1
/teqL
-1
-1
Analytical laboratory assigned tag for Mg (U or M).
Dissolved manganese measured by inductively coupled plasma. The
-1
detection limit was 2/ig L .
Analytical laboratory assigned tag for Mn (U or M).
Dissolved sodium measured by inductively coupled plasma. The
detection limit was 0.87 pec\ L"1.
Analytical laboratory assigned tag for Na (U or M).
Dissolved ammonia-N measured by an auto-analyzer using the
phenate colorimetric method. The detection limit was 0,36 ^eq L"1.
Ammonium ion was estimated for charge balance using:
NH4ION = NH3(1Q-pH)/[5.012 x lO'10 + (10'pH)].
This value differed little from NH3 due to the circumneutral pH of the
lakes.
Analytical laboratory assigned tag for NH3 (U or M).
Dissolved nitrate concentration measured by ion chromatography.
The detection limit was 0.36 fieq L"1.
Analytical laboratory assigned tag for NO3" (U or M).
The organic ion concentration, as calculated using the Oliver model
(Oliver et al., 1983) from DOC and pH using the following equation:
ORGION = (KONST)(DOC)(10)/(KONST+10'pH).
where KONST = 10-(-96+-9(pH>-°-039(pH^
Dissolved phosphorus measured by single reagent ascorbic acid
automated colorimetry, after ammonia-persulfate digestion. The
detection limit of this method was 7 fig L"1.
Closed-cell pH, measured according to U.S. EPA (1987). Samples
were taken in syringes directly from the Van Dorn sampler, and
delivered to pH measurement cells from the same, without intro-
20
-------
PH LAB
P_J
SAMPCODE
SAMPLE N
SCO3
SECCHI m
SI
-1
-1
ducing air to the sample. This pH measurement is more accurate
than the laboratory pH, as holding times for the laboratory pH
measurements ranged up to three months.
pH measured by the analytical laboratory, as late as three months
after sample collection. These pH measurements were taken from
sample bottles in open containers, and were thus subject to gas
exchange.
Analytical laboratory assigned tag for P (U or M).
Indicates whether an observation is a Routine or QA sample.
KLIPDS3 contains all samples—routine, duplicate, audits, blanks, and
splits. KLIPDS4 contains only routine samples, and where duplicates
were available, averages of the duplicate pairs.
R = Routine
D = Duplicate
BFIL = Filtered Blank
BFLD = Field Blank
A = Audit, unfiltered
AFIL = Audit, filtered
S = Split
This is a unique identifier from the analytical laboratory. In cases
where the data were averaged in KLIPDS4, the number refers to only
one of the duplicate pair.
The concentration of carbonate and bicarbonate estimated from the
ANC and pH measurements as follows:
SCO3 = ALKJOT - OHMINUS + HPLUS,
where OHMINUS = (10
_ HnPHSTVL-1
4)(106).
(When SCO3 < 0, SCO3 was set equal to 0.)
The Secchi disk depth, taken as the mean of depths of
disappearance and reappearance.
Dissolved silicon concentration measured by inductively coupled
plasma, with a detection limit of 40 fig L"1.
21
-------
SI_T
SITE DP m
SO4
SO4JT
VISITED
-1
Analytical laboratory assigned tag for Si (U or M).
The depth of the lake at the sampling site in meters. An attempt was
made to sample at the deepest part of the lake located by sonar, so
this measurement serves as an estimate of maximum lake depth as
well.
Dissolved sulfate concentration measured by ion chromatography.
Analytical laboratory assigned tag for SO42".
Not all of the selected lakes were sampled. Some were dry and
others had nesting swans, so we were denied access. Location
information about lakes Not Sampled, or Not Visited exists in
KLIPDS3 but not in KLIPDS4, although the variable VISITED occurs in
both datasets.
WEIGHT
WSAREA ha
WTEMPB
WTEMPS
VS = lake was visited and sampled.
NV = lake was not visited. This occurred due to nesting swans
on the lake. We refrained from sampling in order not to
disturb the nests.
VNS = lake was visited but not sampled. This occurred twice, due
to wetland areas portrayed as lakes on the map.
This is the statistical weighting factor to be applied to each lake when
population estimates are desired. As the sample was randomly
drawn from a defined population, each lake sampled represents a
specified proportion of the population. The sample was not stratified,
so weighting factors for all lakes are the same.
The watershed area of the lake, in hectares, as measured on
1:25,000-scale USGS topographic maps.
The water temperature measured 1.5 m above the bottom of the lake.
The water temperature measured at 1.5-m depth. A difference of
> 4°C between surface and bottom temperatures indicated thermal
stratification of the lake.
22
-------
6. ASCII DATASET FORMAT
This section describes the files and information contained on the KLIP data diskette. The
following files are included:
KLIPDS4.ASC ASCII listing of KLIPDS4, suitable for input to computer program.
KLIPDS3.ASC
FORMAT.TXT
CONTENTS.LIS
KLIP.DAT
ASCII listing of KLIPDS3, suitable for input to computer program, same
format as KLIPDS4.ASC.
Format description for KLIPDS4.ASC and KLIPDS3.ASC, suitable for
translation to various programming languages.
A list of variables and their SAS labels, which are short definitions of
variables, including variable units.
A SAS transport file (Version 6.04, created using PROC CPORT) of
KLIP Dataset 3 (KLIPDS3.SSD) and Dataset 4 (KLIPDS4.SSD).
FORMAT.TXT, shown on the next page, describes the format for the KLIPDS3 and KLIPDS4
ASCII files. The format of this file is the one used to input the data into a SAS data file. However,
the structure is set up for easy editing, so that the file can be input into other databases as well.
The listing of FORMAT.TXT is followed by CONTENTS.LIS, a list of the variables included, with
brief definitions, including the units for each variable in the dataset. This is the list that would be
created by running a PROC CONTENTS on the SAS datasets. PROC CONTENTS is a SAS
procedure that provides a description of the dataset, including the number of observations, and
the listing of the variables included, along with short definitions included in the variable label.
6.1 FORMAT.TXT
Format description:
1) Record Structure
Each record of the KLIP database spans 10 'cards' or input lines
in files KLIPDS3.ASC or KLIPDS4.ASC.
Input lines are referred to as 'CARD n' in the description below.
2) Field Structure
©50 VARNAME 11.5
I Field width (numeric), decimal width.
Variable Name as listed in file CONTENTS.LIS.
Starting column on CARD n
©50 VARNAME
$10.
k.
rield width (character).
Variable Name as listed in file CONTENTS.LIS.
Starting column on CARD n
23
-------
3) Hissing values are coded as -999 for numeric variables, and as blanks
for character variables in the ASCII files.
4) Card Descriptions
CARD 1
©0 LAKE ID $10.
©12 LAKENAME $25.
©39 AL 11.5
©52 AL T $2.
©56 ALRJTOT 11.5
•
6ARD 2
©0 ANDEF 11.5
©13 HPLUS 11.5
©26 ANSUM 11.5
©39 CA 11.5
©52 CA T $2.
©56 CATSUM 11.5
•
CARD 3
©0 CL 11.5
©13 CL T $2.
©17 COEVAL 11.5
©30 CONDCAL 11.5
©43 CONDFLD 11.5
©56 CONDLAB 11.5
•
6ARD 4
©0 DATE DATE7.
©13 DIS REF 11.5
©26 DOC~11.5
©39 DOC T $2.
©43 DS UOAST 11.5
©56 ELEV 11.5
CARD 5
©0 FE 11.5
©13 FE T $2.
©17 K T1.5
©30 K T $2.
©34 LAKE ACC $6.
©42 LAKE~TYP $8.
©52 LAREA 11.5;
•
CARD 6
©0 LAT DMS $10.
©12 LOHG DMS $10.
©24 HG 1T.5
©37 MG T $2.
©41 MN~11.5
©54 HN T $2.
©58 NA 11.5
©71 NA_T $2.
•
6ARD 7
©0 NH3 11.5
©13 NH3 T $2.
©17 N03~11.5
©30 N03 T $2.
©34 ORGTON 11.5
©47 P 11.5
©60 P T $2.
24
-------
CARD 8
©0 PHSTVL 11.5
©13 PH LAB 11.5
©26 SAMPCODE $4.
©30 SAMPLE_N $10.
©42 SC03 11.5
©55 SECCHI 11.5
CARD 9
©0 SI 11.5
©13 SI T $2.
©17 SITE_DP 11.5
©30 S04 11.5
©43 S04_T $2.
©47 VISITED $3.
©52 WEIGHT 11.5
CARD 10
©0 WSAREA 11.5
©13 WTEMPB 11.5
©26 WTEMPS 11.5
©39 COMMENT $60.
6.2 CONTENTS.LIS
KLIP Variable List
NUMBER OF OBSERVATIONS: KLIPDS3 95; KLIPDS4 59
NUMBER OF VARIABLES: 62
ALPHABETIC LIST OF VARIABLES AND ATTRIBUTES
VARIABLE
AL
ALKJTOT
ALJ
ANDEF
ANSUM
CA
CATSUM
CAJT
CL
CL_T
COLVAL
COMMENT
CONDCAL
CONDFLD
CONDLAB
DATE
TYPE
NUM
NUM
CHAR
NUM
NUM
NUM
NUM
CHAR
NUM
CHAR
NUM
CHAR
NUM
NUM
NUM
NUM
LENGTH
8
8
2
8
8
8
8
2
8
2
8
60
8
8
8
8
LABEL
Total aluminum (ag L"1j
ANC (aeq L'1)
Aluminum tag
Anion deficit (aeq L"1)
Sum of anions (aeq L"1)
Ca (ueq L"1)
Sum of cations (aeq L"1)
Calcium tag
Cl (aeq L"1)
Chloride tag
True color
Comment about observation
Calculated conductance (aM cm"1)
Field conductance (aM cm"1)
Laboratory conductance (aM cm"1)
Sample Date
25
-------
D1S_REF
DOC
DOC_T
DS_COAST
ELEV
FE
FE_T
HPLUS
K
K_T
LAKENAME
LAKE_ACC
LAKEJD
LAKE_SIZ
LAKE_TYP
LAT_DMS
LONG_DMS
MG
MG_T
MN
MNJT
NA
NA_T
NH3
NH3_T
N03
N03JT
ORGION
P
PHSTVL
PH_LAB
P_T
SAMPCODE
SAMPLE_N
SCO3
SECCHI
SI
SITE_DP
SI_T
SO4
S04_T
VISITED
NUM
NUM
CHAR
NUM
NUM
NUM
CHAR
NUM
NUM
CHAR
CHAR
CHAR
CHAR
NUM
CHAR
CHAR
CHAR
NUM
CHAR
NUM
CHAR
NUM
CHAR
NUM
CHAR.
NUM
CHAR
NUM
NUM
NUM
NUM
CHAR
CHAR
CHAR
NUM
NUM
NUM
NUM
CHAR
NUM
CHAR
CHAR
8
3
2
8
8
8
2
8
8
2
25
6
10
8
8
10
1°
8
2
8
2
8
2
8
2
8
2
8
8
8
8
2
4
10
8
8
8
8
2
3
2
3
Distance from refinery (km)
DOC(mgL"1)
DOC tag
Distance from coast (km)
Lake elevation (m)
Iron (ug L"1)
Iron tag
Hydrogen ion concentration (aeq L"1)
K (aeq L"1)
Potassium tag
Lake name
Lake access method
Lake identifier
Lake surface area (ha)
Lake hydrologic type
Latitude (dms)
Longitude (dms)
Mg (aeq L"1)
Magnesium tag
Manganese (ug L"1)
Manganese tag
Na (aeq L"1)
Sodium tag
NH3 (ueq L'1)
NH3 tag
NO3 (aeq L'1)
NO3 tag
Organic anion (aeq L"1)
Phosphorous (ag L"1)
Closed-cell pH
Laboratory pH
Phosphorous tag
Sample type
Sample number
SCO3 calc w/phstvl (aeq L"1)
Secchi depth (m)
Silicon (ug L"1)
Lake depth at sample site (m)
Silicon tag
SO4 (aeq L"1)
Sulfate tag
Lake visitation status
26
-------
WEIGHT
WSAREA
WTEMPB
WTEMPS
NUM
NUM
NUM
NUM
8
8
8
8
Population weighting factor
Watershed area (ha)
Temperature bottom (°C)
Temperature surface (°C)
7. REFERENCES
Aspila, K.I., ed. 1989. A Manual for Effective Interlaboratory Quality Assurance. NWRI Contri-
bution 89-99. National Water Research Institute, Canada Centre for Inland Waters, P.O. Box
5050, Burlington, Ontario, Canada L7R 4A6.
Chaloud, D.J., J.M. Nicholson, B.P. Baldigo, C.A. Hagley, and D.W. Sutton. 1987. Handbook of
Methods for Acid Deposition Studies: Field Methods for Surface Water Chemistry.
EPA/600/4-89/020. U.S. Environmental Protection Agency, Washington, D.C.
Eilers, J.M., D.H. Landers, A.D. Newell, M.E. Mitch, M. Morrison, and J. Ford. In press. Major ion
chemistry of lakes on the Kenai Peninsula, Alaska. Can. J. Fish. Aquat. Sci.
Hogg, R.V., and E.A. Tanis. 1983. Probability and Statistical Inference. Macmillan Publishing Co.,
Inc., New York.
Kaufmann, P.R., A.T. Herlihy, J.W. Elwood, M.E. Mitch, W.S. Overton, M.J. Sale, J.J. Messer, K.A.
Cougan, D.V. Peck, K.H. Reckhow, A.J. Kinney, S.J. Christie, D.D. Brown, C.A. Hagley, and
H.I. Jager. 1988. Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern
United States. Volume I: Population Descriptions and Physico-Chemical Relationships.
EPA/600/3-88/021 a. U.S. Environmental Protection Agency, Washington, D.C. 397 pp.
Landers, D.H., J.M. Eilers, D.F. Brakke, W.S. Overton, R.D. Schonbrod, R.E. Crowe, R.A. Linthurst,
J.M. Omernik, S.A. Teague, and E.P. Meier. 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, D.C. 136 pp.
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
27
-------
Descriptions and Physico-Chemical Relationships. EPA-600/4-86/007a. U.S. Environmental
Protection Agency, Washington, D.C. 275 pp.
Little, R.J., and D.B. Rubin. 1987. Statistical Analysis with Missing Data. John Wiley & Sons,
New York. 61 pp.
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.
SAS. 1988. SAS Language Guide for Personal Computers, Version 6.03. SAS Institute, Inc., Box
8000, Gary, North Carolina 27511-8000. 559 pp.
U.S. Environmental Protection Agency (EPA). 1979. Methods for Chemical Analysis of Water and
Wastes. EPA/600/4-79/020. Environmental Monitoring and Support Laboratory, Cincinnati,
Ohio. ;
U.S. Environmental Protection Agency (EPA). 1987. Handbook of Methods for Acid Deposition
Studies. Laboratory Analyses for Surface Water Chemistry. EPA/600/4-87/026. Office of
Research and Development, Washington, D.C.
U.S.Environmental Protection Agency (EPA). 1992. Data User's Guide to the United States
Environmental Protection Agency's Long-Term Monitoring Project: Quality Assurance and
Data Dictionary. EPA/600/3-91/072. U.S. EPA Environmental Research Laboratory,
Corvallis, OR.
28
-------
APPENDIX A
KLIP SAMPLE LAKES
This appendix contains a list of the 59 sampled lakes that represent the Kenai target population,
their location, and their major ion chemistry.
29
-------
CM
•5
0>
8?
i
s
I
I
O
t
a
07
(8
I
0)
c
o
E
a>
.e
O
_£
lu
o
I
o
o*
CO
CO
a.
II
vvvvvvvvvvvv vvvvvvvvv vvvvvvv
COCpCO*~CVJ^r— lOlOCOCOOOT-IOCO^— ^*T— IO O ^ CM IO CO i— CM O> ^— t— T- •*— t— *—
CM CM ^— ^— v CM
(ddddT^dddddddddddddddddddT^dddddo'd
VVV VV VVVVV V V VVVV V VVVVV
VVVVVVVVVVVVVVVV VV'V'^VVVVV'VVVVVV
co T" t^ i^ l^ r^ T~ oo o co co o o T~ o oj t^ cw w o ^ o co ^ o o co ^ co to co ^
^•COO^— IOCVJO>COO)CMCM.^J*OCMIV«*^— CMOIOCOlOCOOlOOOOh^COOCOOJO)
ES88S89S5S Site
cocM$cooono>l^-(oeoc\iioh-
COCM.OCMOT-IOi-'.-'J-IOlO*
S ozzzzzzz
z z z z z:
30
-------
CM
•5
CM
o
I
8
o>
i
Q.
UJ
T3
£
a.
ts
CO
(0
I
'5
o
o>
£
(0
I
6
c
i
5*
o
•»
8
X
£
S5
jl
1,
S-s-
E £.
-S
J,
oq o co\t^ coqco^wcQiowajco^co'frcocqcq'ircqcoqqiO'*
•(r^^CM^iq'a-'*'*'*;^'*'*'*'*'*-*'*^^;'*'*;'*;'*'*'*-*
dddT-'dcNJdddddo'ddddddddddddddd
VVV V VVVVVVVVVVVVVVVVVVVVV
8 5 8 8
s ^ - -
COIOCOM-COr-CMT-OO
•^IJCOg^T-^CO'r-T-
o'ddddddddddddddd-^dddddddddd
VV VVV VVV V V
3 8 S
8
O>^O)O>in
T— OjT-COT-T—Tj'f**
cqi^cooi^tpcqqcn
i^raoWT-CNjcO'SFto
S
Illc^
8
co'cdco'Kcor^i^cdr^r^t^cdr^r^r^r^i^i
18
!o?5888^^5
SCO CO
^r ^r
COCOCOCOCOCOCDCOCO
I
zzz
^zzJSzzzz
00000000^000000000030
I
31
-------
APPENDIX B
REFERENCE VALUES FOR LRTAP SAMPLES
This appendix contains the reference values for the nine Long Range Transport of Airborne
Pollutants Project audit samples from Study 16, used during the Kenai Lakes Investigation Project
to assess accuracy of the data.
-------
Reference Values for LRTAP Samples
04
ANC
fieq L'1 -43.4
Ca
fieq L-1
Cl
fieq L'1
Color
PCU
Conductance
fiS cm"1
DOC
mg L"1
K
fteq L"1
Mg
^ueq L"1
Na
/teq L"1
75.8
26.7
1.0
36.3
0.5
3.6
28.0
13.5
05
-40.0
88.8
5.6
30.0
36.0
5.7
3.6
33.2
23.5
10
0.00
2138.2
2992.0
1.0
445.0
0.14
23.
794.
839.6
LRTAP
01
-1.4
14.9
14.1
2.0
6.9
0.4
4.1
4.1
14.8
Sample
03
110.9
172.2
119.3
1.0
45.3
0.2
7.7
58.4
145.3
Number
08
50.7
224.5
306.2
30.0
68.0
5.2
20.2
60.1
264.5
07
0.8
109.3
20.6
37.2
26.9
5.9
5.1
45.3
34.4
06
2.0
37.9
118.4
110.0
28.7
10.9
7.4
37.0
136.2
09
814.8
654.6
34.4
5.0
94.0
1.4
12.9
227.9
56.1
'1
'1
PH
Si
mg L
1
25.4 0.5 0.6 0.7 2.1 0.64 0.5 1.3 0.4
53.9 0.6 2.1 16.8 0.8 15.7 2.9 2.5 21.1
4.40 4.42 5.56 5.45 6.99 6.41 5.20 5.11 7.75
0.10 2.4 0.01 0.02 0.16 2.5
2.1
1.2 1.1
110.1 171.1 782.0 7.5 156.5 153.9 143.6 39.3 64.5
33
-------
------- |