4>EPA
United States
Environmental Protection
Agency
Office of Acid Deposition,
Environmental Monitoring and
Quality Assurance
Washington DC 20460
EPA/600/8-90/055
July 1990
Research and Development
National Stream Survey
Database Guide
-------
SUBREGIONS OF THE NATIONAL STREAM SURVEY-PHASE I
Northern
Appalachians (2Cn)
Valley and Ridge (2Bn)
Southern Blue Ridge (2As)
(Pilot Study)
Poconos/Catskills (ID)
NY\
Ozarks/Ouachitas (2D)
Mid-Atlantic
Coastal Plain (3B)
Southern Appalachians (2X)
-------
EPA/600/8-90/055
NATIONAL STREAM SURVEY DATABASE GUIDE
by
Mark E. Mitch
Philip R. Kaufmann
Alan T. Herlihy
W. Scott Overton
Michael J. Sale
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, OR 97333
Environmental Monitoring Systems Laboratory - Las Vegas, NV 89193
-------
NOTICE
The research described 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.; Contract Nos. 68-02-3889 and 68-02-3994 to Radian
Corporation; Contract Nos. 68-03-3246 and 68-C8-0006 to NSI Technology Services, Inc.; Interagency
Agreement Nos. #DW89931368 and #1824-1557-A1 with the U.S. Department of Energy (Contract No.
DE-AC05-84OR21400 with Martin Marietta Energy Systems, Inc.); and Cooperative Agreements with
Utah State University (CR815168) and Oregon State University (CR813061). It has been subject to
the Agency's peer and administrative review, and 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.
The correct citation for this document is:
Mitch, M.E., P.R. Kaufmann, A.T. Herlihy, W.S. Overton, and M.J. Sale. 1990. National Stream
Survey Database Guide. EPA/600/8-90/055. U.S. EPA Environmental Research
Laboratory, Corvallis, Oregon. 92 pp.
Inquires regarding the National Stream Survey - Phase I Mid-Atlantic and Southeast should
be directed in writing to:
Chief, Watershed Branch
U.S. Environmental Protection Agency
Environmental Research Laboratory
200 SW 35th Street
Corvallis, Oregon 97333
-------
RELATED DOCUMENTS
Supplemental information on the National Stream Survey - Phase I (NSS-I) can be found in a
series of ancillary manuals and reports. These publications include:
A Sampling Plan for Streams in the National Surface Water Survey. 1985. Technical Report
114 (July 1986). Overton, W.S. Department of Statistics, Oregon State University,
Corvallis, Oregon 97331. 18 pp.
Draft Research Plan, National Surface Water Survey: National Stream Survey, Mid-Atlantic
Phase I and Southeast Screening. 1985. U.S. Environmental Protection Agency, Office of
Research and Development, Washington, D.C. 134 pp.
Eastern Lake Survey - Phase II, National Stream Survey - Phase I: Processing Laboratory
Report-1988. Arent, L.J., M.D. Morison, and C.S. Soong. EPA/600/4-88/025. U.S.
Environmental Protection Agency, Washington, D.C. 86 pp.
National Surface Water Survey: National Stream Survey, Phase I - Pilot survey. 1986. Messer,
J.J., C.W. Ariss, J.R. Baker, S.K. Drouse, K.N. Eshleman, P.R. Kaufmann, R.A. Linthurst,
J.M. Omernik, W.S. Overton, M.J. Sale, R.D. Schonbrod, S.M. Stambaugh, and J.R.
Tuschall, Jr. EPA/600/4-86/026. U.S. Environmental Protection Agency, Washington, D.C.
179pp.
Quality Assurance Plan for the National Surface Water Survey, Stream Survey (Middle Atlantic
Phase I, Southeast Screening, and Middle Atlantic Episodes Pilot). 1986. Drouse, S.K.,
D.C. Hillman, LW. Creelman, and S.J. Simon. EPA/600/4-86/044. U.S. Environmental
Protection Agency, Washington, D.C. 215 pp.
Field Operations Report, National Surface Water Survey, National Stream Survey, Pilot Survey.
1987. Knapp, C.H., C.L Mayer, D.V. Peck, J.R. Baker, and G.J. Filbin. EPA/600/8-87/019.
U.S. Environmental Protection-Agency, Washington, D.C. 110 pp.
Evaluation of Quality Assurance and Quality Control Sample Data for the National Stream
Survey (Phase I - Pilot Survey). 1987. Drouse, S.K. EPA/600/8-87/057. U.S.
Environmental Protection Agency, Washington, D.C. 45 pp.
Analytical Methods Manual for the National Surface Water Survey, Stream Survey (Middle
Atlantic Phase I, Southeast Screening, and Middle Atlantic Episodes Pilot). 1987.
Hillman, D.C., S.H. Pia, and S.J. Simon. EPA/600/8-87/005. U.S. Environmental
Protection Agency, Washington, D.C. 265 pp.
A Sampling and Analysis Plan for Streams in the National Surface Water Survey. 1987.
Technical Report 117. Overton, W.S. Department of Statistics, Oregon State University,
Corvallis, Oregon 97331. 50 pp.
Data Management and Analysis Procedures for the National Stream Survey. 1990. Sale, M.J.
(editor). ORNL/TM. Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831.
(Draft).
National Surface Water Survey: National Stream Survey (Phase I, Southeast Screening, and
Episodes Pilot). Field Operations Report. 1988. Hagley, C.A., C.L. Mayer, and R.
Hoenicke. EPA/600/4-88/023. U.S. Environmental Protection Agency, Washington, D.C.
36 pp.
HI
-------
National Stream Survey - Phase I Quality Assurance Report. 1988. Cougan, K.A., D.W. Sutton,
D.V. Peck, V. Miller, J.E. Pollard, and J. Teborg. EPA/600/4-88/018. U.S. Environmental
Protection Agency, Washington, D.C. 205 pp.
Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern United States.
Volume I: Population Descriptions and Physico-Chemical Relationships. 1988.
Kaufmann, P.R., AT. 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. EPA/600/3-88/021 a. U.S. Environmental Protection Agency,
Washington, D.C. 397 pp.
Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern United States.
Volume II: Streams Sampled, Descriptive Statistics, and Compendium of Physical and
Chemical Data. 1988. Sale, M.J., P.R. Kaufmann, H.I. Jager, J.M. Coe, K.A. Cougan, A.J.
Kinney, M.E. Mitch, and W.S., Overton. EPA/600/3-88/021 b. U.S. Environmental
Protection Agency, Washington, D.C. 595 pp.
IV
-------
TABLE OF CONTENTS
Section Page
Notice ii
Related Documents iii
Acknowledgements viii
1. INTRODUCTION 1
1.1 Purpose and Scope 1
1.2 NSS-I Database Distribution Notes 2
1.2.1 Description of NSS-I and Pilot Survey Data Sets 2
1.2.1.1 NSS-I Report Data 2
1.2.1.2 Data Set NSSIDS3 4
1.2.1.3 Data Set NSSIDS4 4
1.2.1.4 Data Set SBRYN 5
1.2.1.5 Data Set NSSFSO 5
1.2.1.6 NSS-I Pilot Report Data 5
1.2.1.7 Data Set PILOTDS3 6
1.2.1.8 Data Set PILOTDS4 6
1.2.2 Transfer Media 7
1.3 Notice of Caution and Issues of Interest to NSS Data Users 7
1.4 Episode Pilot Survey 8
2. SURVEY DESIGN 9
2.1 Objectives 9
2.2 Target Population 9
2.3 Sample Reach Selection 11
2.3.1 Stage 1 12
2.3.2 Stage II 12
2.3.3 Effective Sample Size (ESS) 13
2.3.4 Special Interest Sites 13
2.4 Data Collection 15
3. DATABASE DEVELOPMENT 19
3.1 General 19
3.2 Database Evolution and Review 19
3.3 NSS-I Data Qualifiers: Flags 21
3.4 Enhanced Data (Data Set 4) 21
3.5 Selection and Use of Data Sets 22
3.5.1 Data Set 3 Versus Data Set 4 23
3.5.2 The NSS-I Pilot Survey: The Southern Blue Ridge Province 23
4. DEFINING THE TARGET POPULATION 25
4.1 Evaluation Process 25
4.2 Identifying Noninterest Observations and Sites 25
4.2.1 Episode Identification 26
4.2.2 Acid Mine Drainage 26
4.3 Classification of noninterest sites in the NSS-I Database 26
4.4 Drop Code Values 28
5. DATABASE APPLICATION 29
5.1 Generating the Database of Index Values From the Reaches Sampled 29
5.2 Extrapolation to the Target Population 29
-------
TABLE OF CONTENTS (continued)
5.2.1 Sample Weightings for Population Estimates 29
5.2.2 Estimating the Target Population 32
5.2.3 Variance Estimates and Confidence Bounds 33
5.3 Describing the Target Population 34
5.3.1 The Cumulative Distribution Function (CDF) Curve 35
5.3.2 Length Estimates 38
5.3.2.1 First Approach: Length Estimates Based
on Node Chemistry 38
5.3.2.2 Second Approach: Length Estimates Based
on Interpolated Length 38
NSS Database Guide References 42
Appendix A - Notes of Caution and Issues of Interest to Data Users 45
Appendix B - NSS Data Dictionary 55
Appendix C - Pilot Survey Revisions 76
Appendix D - NSS-I Field Observation Variables 81
Appendix E - NSS Card-Image Format Definition 85
vi
-------
ILLUSTRATIONS
Figure Page
2-1 NSS-I subregions 10
3-1 The NSS-I database development process 20
5-1 Calculation of an example cumulative distribution function curve
for downstream sites 37
5-2 Interpolated length estimates for subregion 2As 39
TABLES
Table Page
1-1 NSS-I Databases Available 3
2-1 NSS-I Grid Point and Effective Sample Size Summary 14
2-2 Variable Measured in the NSS-I 16
5-1 Number of Visits, Total Observations, Target and Nontarget Samples,
and Indexed Observations by Subregion 30
5-2 Total Resource Estimates of the NSS Refined Target Population 36
VII
-------
ACKNOWLEDGEMENTS
We gratefully acknowledge the many people who helped create the NSS-I and Pilot Survey
databases. A project as massive as the NSS could not have been successfully completed without the
cooperation of literally hundreds of individuals, including geographers, statisticians, field and
laboratory crews, QA staff, data base managers, and project scientists.
We thank Allison Pollock (SAI), Joe Bernert (E&S Chemistry), Barry Rochelle (NSI), and Deb
Chaloud (EMSLV) for their reviews of the manuscript.
We would like especially to thank Susan Christie (NSI) for editing and making suggestions for
this document; Jan Coe, Yetta Jager, and Mary Alice Faulkner (SAI), whose data management exper-
tise and programming skills were crucial in bringing the database into existence; Stasia Allen, Doug
Brown, and Jeff Irish (NSI) for their geographic expertise and GIS skills; Karen Cougan, Valerie
Sheppe, Chuck Monaco, and Sevda Drouse (LEMSCO) for their fastidious attention to detail and
quality assurance; Jim Blick (NSI) and Charles Ariss (Utah State University) for their perseverance
through the unending hours of tedium and for their reverence for statistics.
viii
-------
SECTION 1
INTRODUCTION
1.1 PURPOSE AND SCOPE
The National Stream Survey (NSS) is one component of the National Surface Water Survey
(NSWS), a project implemented by the United States Environmental Protection Agency (EPA) as part
of the Aquatic Effects Research Program (AERP). The AERP, which includes several integrated
studies conducted in areas containing surface waters potentially sensitive to change as a result of
acidic deposition, addresses four major policy issues relating to the effects of acidic deposition on
aquatic ecosystems: (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 is not expected, and (5) the rate of change or recovery of aquatic ecosystems if
deposition rates are decreased.
The data contained on the accompanying tape or floppy diskettes were collected during the first
phase of the National Stream Survey (NSS-I), conducted in the mid-Atlantic and southeastern United
States (Kaufmann et al., 1988), and include results of a pilot stream survey conducted in the Southern
Blue Ridge Province (Messer et al., 1986, 1988). Like other components of the U.S EPA's National
Surface Water Survey (NSWS), the NSS is based on a probability sample from an explicitly defined
population of surface waters. Data were collected from these surface waters during what was con-
sidered to be a representative index period. Sample information was then extrapolated to represent a
target stream population within surveyed geographic regions. In the NSS-I, these data were based on
samples collected at upstream and downstream locations on stream reaches during the spring of
1986 (1985 for the Pilot Survey).
The flexibility of the NSS design permits the examination of any subpopulation of the total number
of sampled streams, based on measured attributes. For example, a subpopulation of interest could
be defined as that set of streams located in the state of Maryland or a set of streams with pH values
less than or equal to a particular reference value (e.g., pH <. 5.5).
This document serves as a database guide, providing an overview of various aspects of survey
design, database structure, and statistical applications of the NSS-I and Pilot Survey databases. It
focuses on specific issues that should be kept in mind during analysis and interpretation of the data,
such as the criteria used to identify noninterest sites during the database development process.
An additional document, Data Management and Analysis Procedures for the National Stream
Survey (Sale et al., 1990), details the statistical analysis procedures and computer code listings of
programs used to generate the population estimates presented in Volumes I and II of the NSS-I report
(Kaufmann et al., 1988; Sale et al., 1988).
1 '
-------
This description of the design, analysis, and application of the data has been extracted and sum-
marized from Volumes I and II of the NSS-I report (Kaufmann et al., 1988; Sale et al., 1988). This
document covers the statistical and conceptual design of the surveys (Section 2), the structure and
components of the database (Section 3), and the method by which the sampled stream reaches were
used to describe the NSS-I target population (Section 4). It also discusses estimation of variance and
other statistical issues regarding the computation of regional population estimates (Section 5).
Appendix A discusses specific issues the NSS data user should be aware of. Appendix B
contains a data dictionary that lists variable names and their definitions. Appendix C summarizes
differences in the estimates reported in the NSS-I Pilot Survey report (Messer et al., 1986) and in the
NSS-I final report (Kaufmann et al., 1988). Appendix D lists supplemental variables on which data
were collected in the NSS-I.
Table 1-1 lists the NSS databases available for distribution. The enclosed data files are referred
to as "Data Set 3" and "Data Set 4". Data Set 3 is a verified and validated data set containing the
original data along with observation-specific flags and tags. Data Set 4, a subset of Data Set 3,
incorporates the averaging of field duplicate [quality control (QC)] samples with corresponding routine
water samples and the replacement of values for missing and/or erroneous data identified in a series
of intensive quality assurance (QA) reviews. Data Set 4 (NSSIDS4 and SBRSYN) includes data for
both probability sample and special interest reaches and contains unique identifiers that can be used
to subset the data and calculate population estimates presented in Kaufmann et al. (1988). Data set 4
retains data for multiple observations (e.g., multiple sampling visits).
1.2 NSS-I DATABASE DISTRIBUTION NOTES
1.2.1 Description of NSS-I and Pilot Survey Data Sets
Six data sets are available for distribution (Table 1-1). Four data sets, NSSIDS3, NSSIDS4,
NSSIFSO, and SBRSYN, were used to make the NSS-I estimates presented in Kaufmann et al., (1988).
The two remaining data sets, PILOTDS3 and PILOTDS4, are associated with estimates presented for
the NSS-I Pilot Survey report (Messer et al., 1986).
1.2.1.1 NSS-I Report Data--
Data sets NSSIDS3, NSSIDS4, and SBRSYN were used for data analyses and presentations
made by Kaufmann et al. (1988) and Sale et al. (1988). These data sets contain data collected during
field activities conducted in the spring of 1986 in portions of the mid-Atlantic and southeastern United
States, in addition to data from the Southern Blue Ridge Province, which was revised after the original
-------
TABLE 1-1. NSS-I DATABASES AVAILABLE
CO
Number of
Observations
Number of Variables
Flags/Tags Present
Missing/Erroneous
Data
Duplicate Samples
Approximate Size
Unique Identifiers
NSS Subregion
NSS-I
Data Set 3
(Validated)
NSSIDS3
1487
117
Yes
Present
Retained
2.0Mb
BAT ID with
SAMJD
All
NSS-I
Data Set 4a
(Enhanced)
NSSIDS4
1765
94
No
Substituted
Averaged
t.8Mb
STRM ID with
SAMRN
All
Pilot Survey
Data Set 3
(Validated)
PILOTDS3
397
107
Yes
Present
Retained
0.5Mb
BAT ID with
SAMJD
2As
Pilot Survey
Data Set 4
(Enhanced)
PILOTDS4
339
86
No
Substituted
Averaged
0.3Mb
STRM ID with
SAMRN
2As
Pilot Survey
Synthesized
Datab
SBRSYN
34
61
No
N/A
N/A
0.03Mb
STRM ID
with SAMRN
2As
Field Site
Observations
NSSFSO
1068
73
N/A
N/A
N/A
1.3Mb
STRM ID
with DATSMP
All
a Data for the NSS Pilot Survey reaches have been appended to NSSIDS4. Data comparable to the full-scale survey's spring index baseflow
period are encoded as SAMRN = 1-3 in the Pilot Survey.
b All chemistry data contained in this data set were calculated. These values were used to provide estimates of chemistry for 34 upstream
sampling sites not sampled during the spring baseflow period in the Pilot Survey. In order to uniquely identify these observations and still
incorporate them in to the spring index calculations all 34 of these enhanced observations have a sample visit number (SAMRN) value of 1.5
(see Section 1.2.2.4).
Note: NSS-I population estimates were made using Data Set 4, after averaging multiple site visits (see Section 5.1 for detailed instruction).
-------
Survey report was published (see Section A. 11). The NSS Phase I survey was a broader regional
application of methods developed during the Pilot Survey. In the mid-Atlantic, 250 stream reaches
were visited twice at upstream and downstream sampling locations. In the Southeast, 200 stream
reaches were visited once at upstream and downstream sampling locations. These sample visits are
encoded in the database as SAMRN = 1 or 2. Each sample site is identified with a 9-character
stream identification number comprised of an 8-digit reach ID and a sample site position (U =
upstream; L = downstream). The data from field sampling crews, the processing laboratory, and
analytical laboratories have been merged into one file. In contrast, Pilot survey sampling sites are
designated by 8-character stream identification codes.
1.2.1.2 Data Set NSSIDS3--
Data set NSSIDS3 contains NSS-I data without any type of enhancement or substitution. The
data have undergone an intensive verification and validation process. Data from field duplicate (QC)
water samples are included in the data set as separate observations. Values that have been identified
as suspect, erroneous, or missing are flagged in the data set but are not replaced. Data qualifiers,
flags (see Section 3.3), are included for each chemical variable to denote suspect values, as well as
specific conditions or circumstances pertaining to individual water samples (e.g., holding time
violations). The parameters are presented in the units and precision in which they were originally
measured (see Appendix B, Database Dictionary).
1.2.1.3 Data Set NSSIDS4--
Data set NSSIDS4 is considered the final data set and is the end product of intensive quality
review. This data set is used to generate population estimates. Based on chemical relationships
within the data, erroneous and missing values have been replaced with estimated values (see Section
3.4). Chemical values from field duplicate (QC) water samples have been averaged with correspon-
ding routine samples. After the original Pilot Stream Survey Report was published, a small portion of
the data was revised. The Pilot Survey data, with all revisions, are included in NSSIDS4. The resulting
data set contains observations for 450 NSS-I probability sample reaches, 54 Pilot Survey reaches, and
44 special interest streams. Data are not averaged between sample visits. Note that the reaches
sampled during the Pilot Survey have 8-character stream identification codes, whereas reaches
sampled during the NSS-I have 9-character codes. Sites considered noninterest in generating esti-
mates of the NSS-I target population of streams are identified by a variable (DRPCDE) that contains a
sample exclusion 'drop' code: 0-5 (see Section 4.4).
-------
1.2.1.4 Data Set SBRSYN--
In the Pilot Survey, only 20 of the 54 probability reaches were sampled at both upstream and
downstream sampling locations during the spring index base flow period. A supplemental data set
(SBRSYN) was synthesized to provide spring upstream estimates for the Southern Blue Ridge
Province compatible with data from other NSS-I subregions. Data for 22 chemical variables for the 34
upstream sites that were not sampled were synthesized, based on empirical relationships with data
from the sampled streams and on data collected at all sites during the summer flow period (see
Kaufmann et al., 1988, Appendix B). These observations have been assigned a sample visit number
(SAMRN) of 1.5 to allow the user to identify these sites and merge their data with other Southern Blue
Ridge data that have sample visit numbers 0 - 4. The observations that represent the spring index
period in the Southern Blue Ridge have sample visit numbers 1 - 3. These observations are identified
by a similar 8-character variable, STRMJD, used for Pilot Survey Southern Blue Ridge sites. In order
to calculate population estimates for the Southern Blue Ridge Province that duplicate those in the
NSS-I final report, these synthesized values should be appended to the NSSIDS4 data set prior to
subsetting and averaging data to make population estimates.
1.2.1.5 Data Set NSSIFSO--
Originally titled 'Watershed Disturbance Characteristics", this data set, NSSIFSO, contains data
that describe the watershed in the immediate vicinity of field sampling sites. This information includes
details on immediate watershed disturbances, bank vegetation cover, stream substrate, and additional
field comments about the sample site. Assessment of substrate type and bank vegetation are made
as percentage estimates as: absent (0%), sparse (< 25%), moderate (25-75%), and heavy (> 75%).
Data for both the Pilot Survey and the Phase I Survey are combined in one data set. This information
is based on unvalidated observations of field crews and is therefore subjective. Vegetation coverage
and substrate composition estimates were based on the crew's judgment and are only rough esti-
mates based on visual assessments. This information has not been subjected to a stringent QA
review similar to that received by the chemical data. Although specific observations, coverage
estimates, and substrate assessments may be difficult to validate, these data can be a useful tool in
examining specific conditions that may have existed at the time of sampling (Appendix D).
1.2.1.6 NSS-I Pilot Report Data-
The Pilot Survey data (PILOTDS3, PILOTDS4) were collected during field activities conducted in
the spring and early summer of 1985 in the Southern Blue Ridge Province, Subregion 2As (Messer et
al., 1986, 1988). Five field sampling visits were made from mid-March to mid-June. The Pilot Survey
-------
was conducted to test the logistical and analytical protocols planned for the full-scale NSS-I in the
Mid-Atlantic and Southeast. Data from the Pilot Survey were used to evaluate the statistical sampling
design, logistics plan, quality assurance plan, data management program, and data analysis plan.
Up to 5 sampling visits were made to 54 probability sites and an additional 7 special interest
sites. The 5 Pilot Survey sampling visits are numbered 0-4 (using the variable SAMRN) in the
database. This includes an initial reconnaissance and methods development visit (coded as '0' in the
database) that is not used in making population estimates but may be of interest for examining
temporal variability. The spring index base flow period is represented by sample visits numbered 1-3.
The data associated with sampling visit 4 were collected during a summer base flow period, but may
be useful for additional analyses. Each Pilot Survey sample site is identified by an 8-character stream
reach identification code made up of a 7-character reach ID and a sample site location code (U =
upstream; L = downstream).
After the Pilot Survey Report was published, a small portion of the data was revised (see
Appendix C). The updated and revised data set has been appended to NSSIDS4. PILOTDS3 and
PILOTDS4 are associated with the Pilot Survey Report and do not contain the revised Pilot Survey
data included in NSSIDS4.
1.2.1.7 Data Set PILOTDS3--
Data set PILOTDS3 contains NSS-I Pilot Survey data without any type of enhancement or substi-
tution. The data have undergone a rigorous verification and validation review. Chemical values from
duplicate water samples are included in the data set as separate observations. Values identified as
suspect, erroneous, or missing are flagged in the data set but are not replaced. Data qualifiers, flags,
are included for each chemical variable to denote suspect values, as well as notes about specific
conditions or circumstances pertaining to individual water samples (e.g., holding time violations).
Parameters are presented in the units in which they were originally measured (Appendix B).
1.2.1.8 Data Set PILOTDS4-
Data set PILOTDS4 is the final data set for the Pilot Survey, and is the end product of intensive
quality review. This data set is used to generate population estimates. Based on chemical
relationships within the data, erroneous and missing values have been replaced with estimated values.
Field duplicate (QC) chemical values have been averaged with corresponding routine samples. Note
that all five of the Pilot Survey sample visits are included in the data set. The spring index base flow
period is represented by sample visits (SAMRN): 1 - 3.
6
-------
1.2.2 Transfer Media
The transfer media on which the NSS-I databases are available include either 9-track magnetic
tape or 51/4 inch high-density floppy disks in a card-image ASCII format or in a SAS (SAS, 1985) format
as a SAS export data set. Missing values are replaced in Data Set 4 for both the Pilot and Phase I
databases. In Data Set 3, missing values are represented with the number -999.000 for card image
formats. Standard SAS notation for missing values is used in all SAS files (i.e., a ".' for numeric
variables and •' or blank for character variables). The card-image format definitions are discussed in
Appendix E.
1.3 NOTICE OF CAUTION AND ISSUES OF INTEREST TO NSS DATA USERS
Use of the NSS data can be very complex. Estimates can be generated for a variety of popula-
tions. The user should understand the following specific issues before performing any analyses of
NSS data. Details and context of these issues are summarized in Appendix A.
• Use of Drop Codes: DRPCDE
• Generating a Working Data Set
• Data Set of Field Site Observations
• Use of Sample Weights: W
• Chemical Variables With Similar Names
• Population Estimates for Geographic Subsets of Streams
• Reach Length Estimates: RCH_LN versus L2
• Topographic Drainage Area Measurements: av 3^, 83, a4, a5, atota| (A1, A2, A3, A4, A5 and
A_WS)
• Synthesized Data for the Southern Blue Ridge (Subregion 2AS)
• Differences Between the NSS-I and the NSS-I Pilot Survey
• Revisions to NSS-I Pilot Survey Data
• Comparison of Parameter Units in the NSS-I and the NLS
• Revisions to a1
• NSS Database Variable Formats
• Subregion Identification Codes
Using DIG and pH in Calculated Variables
-------
1.4 EPISODE PILOT SURVEY
Concurrent with the NSS-I field sampling activity, an Episode Pilot Survey was conducted in the
Mid-Atlantic Region. The primary objective of the Episode Pilot Survey was to assess the feasibility of
a synoptic assessment of precipitation episodes using a probability-based sampling design. Although
based on the NSS-I, the Episodes Pilot Survey was not able to collect an adequate number of 'epi-
sode" samples to make regional estimates of extent, duration, or frequency of episodic conditions in
streams. The primary conclusion of investigators was that this type of synoptic survey of episodes
would not be cost effective (Messer and Eshleman, 1987). The data collected as part of the Episodes
Pilot Survey has not been subjected to the same extensive QA review as the NSS-I samples and is not
distributed with the NSS-I data sets.
8
-------
SECTION 2
SURVEY DESIGN
2.1 OBJECTIVES
The NSS-I was designed to chemically and physically characterize a target population of streams,
based on a probability sample. The primary objectives of the NSS-I were:
• To determine the percentage, extent (number, length and drainage area), location, and
chemical characteristics of streams in the United States that are presently acidic, or that
have a low acid neutralizing capacity (ANC) and thus might become acidic.
• To identify streams representative of important classes in each region that might be selected
for more intensive study or long-term monitoring.
The NSS sampling design employed a randomized, systematic technique for selecting a
probability sample of stream reaches within areas of the United States expected to contain waters of
low acid neutralizing capacity (ANC) (Messer et al., 1986; Overton, 1986, 1987; Kaufmann et al., 1988).
Statistically, the stream survey is a double sample, stratified on subregion and expected ANC class (<
50 |ieq L"1 and > 50 jieq L'1), as represented on maps of expected alkalinity prepared by Omernik
and Powers (1983) and Omernik and Kinney (1985).
2.2 TARGET POPULATION
The NSS-I focused on a population of stream reaches defined in terms of size, general water
quality, and location. This target population of reaches, in a broad sense, can be thought of as all
reaches appearing as blue-lines on 1:250,000-scale United States Geological Survey (USGS) topo-
graphic maps in areas expected to contain surface water ANC < 400 (ieq L'1. This population was
restricted to reaches that drain watersheds < 155 km2 and that are not grossly polluted. The primary
intention of surveying reaches in this size range was to examine those streams large enough to be
recreationally and economically important for fish habitat, yet still small enough to be susceptible to
change as a result of acidic deposition. These criteria identified a different set of streams in each
geographic area (e.g., forested upland sites are noticeably different from those located in the Florida
lowlands). The attributes of large bodies of flowing water (i.e., those with Strahler orders > 5 on
1:250,000-scale USGS maps) were not assessed or described in the NSS-I.
The NSS-I was conducted in two regions of the United States, the Mid-Atlantic and the Southeast,
focusing on areas with high sulfate deposition that were considered most likely to contain clear water
-------
SUBREGIONS OF THE NATIONAL STREAM SURVEY-PHASE I
Northern
Appalachians (2Cn)
Valley and Ridge (2Bn)
Poconos/Catskills (ID)
Southern Blue Ridge (2As)
(Pilot Study)
Mid-Atlantic
Coastal Plain (3B)
Ozarks/Ouachitas (2D)
AL *i GA
Wr
Southern Appalachians (2X)
Figure 2-1. NSS-I subregions.
10
-------
streams (low in organic acids) with low alkalinity. Nine subregions were identified within these areas
(Figure 2-1):
1. Poconos and Catskills (Subregion 1D).
2. Mid-Atlantic Coastal Plain (3B), including the New Jersey Pine Barrens and the Chesapeake
Bay area.
3. Northern portion of the Valley and Ridge Province (2Bn).
4. Northern Appalachians (2Cn).
5. Piedmont (3A).
6. Ozarks and Ouachitas (2D).
7. Southern Blue Ridge (2As), NSS-I Pilot Survey study area sampled in 1985.
8. Southern Appalachians (2X), a combined area of portions of the Southern Appalachian
Plateau, the southern area of the Valley and Ridge Province and a northern portion of the
Blue Ridge Mountains.
9. Florida (3C).
Section A. 15 details the identification and identification codes of NSS-I subregions in the NSS
database.
Initially, the NSS-I was divided into two studies: the full-scale Phase I Survey in the Mid-Atlantic
and a screening survey in the Southeast. The principal survey, designed to meet the NSWS primary
and secondary objectives, was conducted in the Mid-Atlantic. In contrast, the Southeast Screening
Survey was designed to assess the need for additional sampling efforts outside the Mid-Atlantic
Region. The difference between these two studies was in the number of water samples collected at
each reach end. In the Mid-Atlantic, two water samples were collected from each stream reach end; a
single water sample was collected for each reach end in the Southeast.
Results from the NSS-I Pilot Survey in the Southern Blue Ridge indicated that a single spring
baseflow sample could adequately represent baseflow conditions and could be subjected to the same
assessment and analysis as the multi-sample Mid-Atlantic Survey (Messer et al., 1986). Based on
these results, data collected from the Mid-Atlantic and the Southeast are considered to make up the
regions sampled in the NSS-I.
2.3 SAMPLE REACH SELECTION
Within the NSS-I, the sampling unit of interest (or description) is the stream reach, defined as a
blue-line segment appearing on 1:250,000-scale maps delimited at each end by a confluence of blue
lines or by the end of a blue line segment (i.e., a headwater reach), conforming to criteria that
excluded noninterest systems (e.g., watershed area > 155 km2, urbanized watersheds, reservoirs or
11
-------
lakes). The NSS-I sample reaches were selected using a stratified two-stage variable probability
process that selectively identified reaches representing the target population of streams in each
subregion. The complete selection process is detailed in Overton (1987).
In contrast to the list sampling frame employed by the National Lake Survey, the NSS-I employed
an area/point frame. A rectangular dot-grid overlay, made of transparent acetate, that projected 64
mi2 per point at a scale of 1:250,000 (i.e., the mapped distance between points horizontally and
vertically was 8 miles) was overlaid on USGS maps. This process identified the Stage I sample, 3,082
blue-line reaches and 222 nonreaches (lakes, reservoirs, swamps). Of the 3,082 potential sample
reaches, 781 were excluded because of drainage area size (> 155 km2) and proximity to urbanized
areas (i.e., more than 20% of the watershed was located within an urban area as indicated on
1:250,000-scale USGS maps). A reach's probability of inclusion at this stage of sampling was directly
related to its direct drainage (a^ and the density of points portrayed on the acetate grid (i.e., 64
midpoint) (Overton, 1987). In order to increase resolution in the high-interest portion of the target
population, a separate stratum of Stage I low ANC sites was identified (those areas where ANC was
expected to be very low, ANC < 50 (ieq L"1).
The Stage II sample, a subset of the Stage I sites, was composed of those sites identified for field
visits (selected as a variable probability sample). AN low ANC Stage I reaches were included in the
Stage II sample (except in the Pilot Survey). Thus the Stage II sample included the low ANC sites and
a probability subsample from the higher ANC stratum. Because all of the Stage I reaches in the low
ANC strata were included in the Stage II sample, these reaches were selected with a Stage II inclusion
probability = 1. In addition, some Stage I reaches had such small a1 values that they also entered
into the Stage II sample with an inclusion probability = 1.
Except for reaches in the low ANC strata and reaches with small a1 values, a reach was selected
in Stage II from the Stage I sample with an inclusion probability inversely proportional to its direct
drainage, a1. The final sample weights, used in making population estimates, are inversely propor-
tional to the overall inclusion probabilities. These weights are equal to the number of reaches each
stream reach represents in the target population. The Stage II sampling process selected 504
reaches for field sampling and chemical measurement.
12
-------
2.3.3 Effective Sample Size (ESS)
A reach's stratum (ANC and subregion) and the effective sample size (ESS) for that stratum are
important values used in calculating variance estimates. A reach's ANC stratum can be identified in
the database by the variable STRATUM. STRATUM is equal to 1, indicating the higher ANC stratum,
to 2, indicating the low ANC stratum, or to 3, indicating a 'small a," stratum. This breakdown of the
number of Stage I and Stage II sites in each stratum is shown in Table 2-1.
Overton (1987) discusses the rationale for calculating the NSS-I ESS. Conceptually, the effective
sample size can be thought of as the number of grid points (interest and noninterest combined) that
would have to be examined in order to obtain a desired number of sample sites. For example, an
effective size of 75 for a subregion indicates that 75 grid points would have to be examined to identify
the desired number of Stage II sites (~ 50 reaches per subregion). For Stage I, this value is the total
number of grid points falling within the boundaries of the study area. Because the Stage II sample is
a subset of the Stage I target sample, the Stage II effective sample size is not explicitly defined.
However, an indirect estimate of this value can be made based on the assumption that the ratio of
target to nontarget sites in the Stage I sample is applicable to the Stage II sample. For the NSS-I
main survey (in contrast to the Pilot Survey1), this value was calculated as:
n" = n2n'/n
where:
• n* is Stage II effective sample size,
• n2 is the Stage II sample size (# of reaches selected for field sampling),
• n' is Stage I effective sample size,
• n is Stage I target reaches (# of Stage I reaches considered targets for Stage II
sampling).
2.3.4 Special Interest Sites
In addition to the reaches selected for the Stage II sample, the NSS-I sampled 36 special interest
reaches. These reaches are similar to those in the NSS-I target population (in terms of size), but are
1The Pilot Survey Stage II selection process chose every other Stage I reach from the list of reaches ordered by grid-point
locations. This is in contrast to the variable probability selection process used in the main survey. In the Pilot Survey, the ESS was
defined as half the number of grid points (84) or the number of points examined to obtain the number of reaches to sample in the
field. Because the main survey used a probability sample to obtain the Stage II sites, the ESS was estimated.
13
-------
TABLE 2-1. NSS-I GRID POINT AND EFFECTIVE SAMPLE SIZE SUMMARY
Stage I Sample
Subregion
1D
2Bn
2Cn
3B
2As
2D
2X
3A
3C
Stratum
(1 = ANC > 50 |ieq L'1)
2 = low ANC)
(3 = small av ANC > 50 neq L"1)
1 or 3
2
1 or 3
2
1 or 3
2
1 or 3
2
1 or 3
1 or 3
1 or 3
1 or 3
1 or 3
#of
Nonreach or
Nontarget
Points
108
1
116
2
93
7
209
7
32
94
91
144
81
#0f
Target
Points
(n)
180
13
380
3
206
19
336
12
136
266
243
423
84
Total
#of
Points
(n')
288
14
4S6
5
279
26
545
19
168
360
333
607
265
Stage II Sample
#of
Selected
Reaches
(n2)
48
13
50
3
55
19
50
12
54
50
50
50
50
#0f
Sampled
Reaches
48
13
50
3
55
19
50
12
54
50
48
50
48
Effective
Sample size8
(n")
76.8
14.0
65.3
5.0
74.5
26.0
81.1
19.0
84.0b
67.7
65.8
71.7
94.3
Effective sample size (ESS) is the number of grid points examined to obtain the needed number of Stage II sample reaches (to be visited in
the field). ESS is stratum specific. Exact values for each ANC stratum are necesary when calculating variance.
The effective sample size is calculated differently in the Pilot Survey than in the full NSS-I survey (see Section 2.3.3).
-------
sites where intensive process-oriented studies or long-term monitoring data are available. Although
field sampling and chemical analyses were undertaken at these sites, they were not used in making
estimates of the target population and their weights are set to zero in the database. These sites,
however, are used to examine watershed processes and long-term trends. The observations for
special interest sites in the NSS database are identified by the following variables:
• SUBJD = "SI1
• DRPCDE = "5-
. w _ .0"
The sixth character in RCHJD and STRMJD = "91
2.4 DATA COLLECTION
The collection of water samples and attributes for the Stage II sample constitutes a third stage of
sampling in the statistical design of the NSS-I. Sampling at this stage characterizes or indexes the
chemical and physical properties of a sample reach. The NSS-I relied on samples taken during an
appropriate season from a representative sample of water bodies to provide an index of the chemical
characteristics of a target population of surface waters (Messer et al., 1986, 1987, 1988). In the NSS,
this index value depicts stream chemistry during spring baseflow between snowmelt and leafout
(approximately March 15 to May 15), when sensitive life stages of important fish species are present
and chemical conditions potentially limiting to aquatic organisms are likely to exist. The rationale of
the spring index sampling period is detailed in Volume I, Section 2.5 of the NSS-I report (Kaufmann et
al., 1988) and in Messer et al. (1988).
Field visits were made during the spring index period to NSS-I reaches in a geographic area of
more than 530,000 km2. The NSS-I water sampling and analysis methodology is detailed in Section 3
of Kaufmann et al. (1988) and in Hillman et al. (1987). On each sample visit, field crews collected a
3.8-L water sample and four 60-mL syringe samples, in addition to recording watershed and hydro-
logic descriptive characteristics and making in situ chemical measurements (Hagley et al., 1988).
Water samples were then transported to a centralized processing laboratory where they were
stabilized. Chemical measurements were made within 36 hours of sample collection (Arent et al.,
1988). The processed samples, aliquoted and preserved, were then shipped to contract analytical
laboratories for chemical analyses. Table 2-2 lists the physical and chemical data collected in the
field, at the processing laboratory, and at the analytical laboratory, as well as the geographic data
recorded for the sample sites.
In addition to routine water samples, QC samples were collected and used to ensure that
sampling and analytical methods were performed according to specifications and to evaluate overall
15
-------
TABLE 2-2. VARIABLES MEASURED IN THE NSS-I
CHEMICAL AND PHYSICAL VARIABLES (see Appendix B)
Parameter
pH in situ
Temperature
Variable
Name
Field Site/In
PH_R
TMPSTR
Parameter
Situ Measurements
Specific conductance
Dissolved oxygen
Variable
Name
CONIS
DOJS
Processing Laboratory Measurements
Monomeric Al
Organic monomeric Al
Dissolved inorganic
carbon (closed system)
Acid neutralizing
capacity (ANC)
Extractable Al
Total Al
Ammonium
Base neutralizing
capacity (BNC)
Calcium
Chloride
Dissolved inorganic
carbon (initial)
Dissolved inorganic
carbon (air equil.)
Dissolved organic
carbon
Fluoride
ALDSVL, ALDS16
ALORVL, ALOR11
ALOR16
DICVAL
Contract Analytical
ALKA11
ALEX11, ALEX16
ALTL11, ALTL16
NH411, NH416
ACCO11
CA11,CA16
CL11.CL16
DICI11
DICE11
DOC11
FTL11.FTL16
Specific conductance
, pH, closed system
True color
Turbidity
Laboratory Measurements"
Iron
Magnesium
Manganese
Nitrate
pH; initial acidity
Titration
pH; initial alkalinity
Titration
pH; air equilibrated
Phosphorus
Potassium
Silica
Sodium
Specific conductance
Sulfate
CONVAL
PHSTVL
COLVAL
TURVAL
FE11, FE16
MG16, MG16
MN11, MN16
NO311, NO316
PHAC11
PHAL11
PHEQ11
PTL11, PTL16
PTD11, PTD16
K11,K16
SIO211, SI0216
NA11, NA16
COND11
S0411,SO416
Multiple variable names exist for those parameters with measurements reported in multiple units (e.g, mg L" and /ieq L"1) or
that have been measured using different methods (see Appendix B).
16
-------
TABLE 2-2. VARIABLES MEASURED IN NSS-I (Continued)
Parameter
Anion deficit
Total anions
Total cations
Carbonate concentration
Bicarbonate
Variable
Name
ANDEF
ANSUM
CATSUM
CO316
HCO316
Parameter
Calculated Variables
Hydroxide cone.
Organic anion cone.
Sum of base cations
Sum of strong mineral acid
anions
Inorganic monomeric
aluminum
Hydronium concentration H16
Variable
Name
OH16
ORGION
SOBC
SOSMA
ALINOR
GEOGRAPHIC VARIABLES
Geographic Attribute
Watershed area contributing to mapped sampling point (km2)
Direct drainage area (mi2) between upstream and downstream
reach ends
Drainage area between upstream/downstream sample sites
Drainage area above upstream sample site
Site elevation (m)
Stream gradient (%)
Site latitude (decimal degrees)
Site longitude (decimal degrees)
Length of reach between upper and lower sampling sites (km)
Length of reach between stream confluences (km)
Name of map(s) showing watershed location
Name of county for reach and watershed
Name of 1:250,000-scale map showing watershed
Number of headwater reaches (Shreve order)
Shreve order
ANC stratum (1 or 2) in statistical design
Strahler order
Strahler order
Reach identification code
Stream identification code
State (2-character code)
Stream name
Subregion identification code
5S Map Scale
1:24,000
1 :24,000
1 :24,000
1 :24,000
1:24,000
1 :24,000
1 :24,000
1:24,000
1:24,000
1:24,000
1:24,000
1:24,000
1:250,000
1:250,000
1:24,000
(section 2.4)
1:24,000
1 :250,000
N/A
N/A
1 :250,000
1 :24,000
N/A
Variable Name
A_WS
A1
A4
A5
ELEV
GRADE
LAT STD
LON STD
L2
RCH LN
MAPI. .6
COUNTY1..4
QUAD
RCH HW
SHREV75
STRATUM
STRA75
STRA250
RCH ID
STRM ID
STATE1
STRMNAM
SUB ID
17
-------
data quality for the survey. The types and uses of these samples are detailed in Cougan et al. (1988),
Drouse et al. (1986), and Kaufmann et al. (1988).
It should be noted that a reach was considered sampled if a visit was made by a field crew to a
mapped sampling point. Though specific stream characteristics and/or watershed conditions might
have eliminated a reach from the NSS target population, all collected information (even for noninterest
conditions) can be used to describe the characteristics of some populations of reaches represented in
the Stage II sample.
18
-------
SECTION 3
DATABASE DEVELOPMENT
3.1 GENERAL
This section provides background information on the database development process used the
NSS-I and Pilot Surveys. An important component of all NSWS projects is the approach used in
database development to ensure that the collected and recorded data are representative of the
physical and chemical characteristics of the water body at the time of sampling. The NSS-I database
contains over 56,000 individual values, including physical and chemical parameters as well as quality
assurance data, all of which were reviewed individually and in the context of subregion chemistry.
Additional details of the database development process and data analysis procedures are presented
in Kaufmann et al. (1988) and Sale et al. (1989). As with all NSWS databases, the NSS-I data has
undergone an external audit (Pollack and Grosser, 1988) to verify and validate data quality
assessments and data set evolution.
3.2 DATABASE EVOLUTION AND REVIEW
The final data sets used in making NSS-I population estimates have been subjected to four levels
of QA evaluation. The completion of each level of QA review produced a new working data set of
greater refinement (Figure 3-1). These working data sets are defined as : raw (Data Set 1), verified
(Data Set 2), validated (Data Set 3), and enhanced (Data Set 4). The final product of this refinement
process, the enhanced data set (Data Set 4), incorporates data substitution and replacement of
missing values. Data Set 4 was used in calculating the NSS-I population estimates presented in
Volume I of the NSS-I report (Kaufmann et al., 1988).
The verification process focused on the internal consistency of chemical measurements within a
water sample. This process identified individual chemical values that appeared as exceptions to
results predicted from calculations based on chemical relationships (e.g., anion/cation balances, con-
ductance estimates, and protolyte analysis). Concurrent with this process, QC samples (including
audit samples, field blank data, and instrument detection limit values) were assessed for potential ana-
lytical bias in the laboratory or the field. The final product from the verification process is Data Set 2.
The validation process examined stream chemistry in the context of the total group of sample
streams within a subregion. Observations that were identified as "atypical" during review of data in a
subregional context were considered outliers from the rest of the data. Sites showing a number of
unusual chemical values generally were not suspected of having serious analytical errors, but rather
were associated with site observations that suggested probable impacts from watershed disturbances,
such as acid mine drainage, tidal influence, or urban runoff. Individual chemical outliers were
19
-------
rield Sampling^
Processing
.Laboratory
Analytical
Laboratories
I
Visual Form Check
Data Entry I
Data Entry 2
Error and
Range Check
RAW DATA SET
(Data Set I)
Batch Reports
Verification
Data Editing
and Flagging
^ I
VERIFIED DATA SET
(Data Set 2)
Site
Reports
Maps
Validation
Data Editing
and Flagging
I
Substitution
and
Replacement
' II
VALIDATED DATA SET
(Data Set 3) |]_
ENHANCED DATA SET
(Data Set 4)
Figure 3-1. The NSS-I database development process.
20
-------
examined for errors that might have occurred during transcription or analysis. Erroneous or missing
values were considered for possible replacement for calculating population estimates. The final
product of the validation process is Data Set 3.
3.3 NSS-I DATA QUALIFIERS: FLAGS
In the development of the NSS-I databases, data qualifiers were used as a tool to mark individual
values or even an entire water sample as having particular features (e.g., sample holding time issues,
analytical instrument errors, QA discrepancies) that could be pertinent to data interpretation. These
qualifiers, described as flag variables, identify observations or notes made during the QA process
(verification and validation) employed in creating the enhanced database. Flag variables help identify
nonrepresentative or questionable data. The NSS-I flags are included in Data Set 3 but not in Data
Set 4 because of the averaging of duplicate samples used in generating Data Set 4 and for ease of
use of the data set.
Each chemical variable has a complementary flag variable with a prefix of the chemical variable
name and an "F' to indicate a flag. For example, the flag variable for ALKA11 is listed under the
variable name ALKA11 F. A complete listing of all flag definitions is presented in the NSS-I Data
Dictionary (Appendix B, Table B-1). A flag code is composed of two characters, an alpha and a
numeric. The first, the alpha character, identifies a problem or concern category (e.g., anion/cation
balance discrepancies). The second, the numeric, indicates a specific problem or note within the
category. For example, if the flag variable MG11F contained the code "A3", this indicates an
anion/cation balance discrepancy with possible cation contamination. Flag variables may also contain
multiple codes. If a variable contained the code "A3BOHO", this would indicate three flagged con-
ditions: (1) a problem with the anion/cation balance, (2) a problem with a field blank water sample for
the batch of samples, and (3) holding time criteria were not met.
3.4 ENHANCED DATA (DATA SET 4)
Data Set 4, a subset of Data Set 3, is used to generate a data set for calculating population
estimates and associated statistics. Such estimates are difficult to generate if there are
inconsistencies in the data (e.g., missing values). Data Set 4 was prepared to resolve problems of
missing and erroneous data. When necessary, substitutions were performed according to the
following protocol:
1. Whenever possible, values from duplicate (QC) water samples were used.
2. If a duplicate measurement was not available, a value from an alternate visit to the site was
used.
21
-------
3. If a duplicate measurement or a measurement from an alternate visit was not available, a
substitution value was calculated by means of a linear regression model or ion balance
estimate. This predicted value was calculated based on observed chemical relationships
and spatial relationships between upstream and downstream sites.
All values generated for substitution were examined for consistency with other data before placing
them in the final data set. Only data from routine and duplicate samples were assessed for substi-
tution or enhancement. Of the 44, 975 values, 83 (< 0.5% of the database) were flagged. Only 34
values were replaced (enhanced) in Data Set 4. Of these 34 values, 20 were replaced with data from
alternate visits. The remaining 14 were replaced using multiple linear regression and ion balances. In
addition to substitutions for erroneous and missing values, negative values for parameters other than
ANC and base neutralizing capacity (BNC) were set equal to zero2. All modified values in the final
data set are flagged.
In roughly 10% of the samples collected, duplicate samples were collected as QC checks to esti-
mate sample precision. The data from these duplicate samples were averaged with data from the
routine samples in generating the enhanced data set.
3.5 SELECTION AND USE OF DATA SETS
The distributed NSS-I and Pilot Survey databases contain chemistry data, geographic attributes,
and watershed descriptive information. For both the NSS-I and the Pilot Survey, three data sets are
available: Data Set 3 (NSSIDS3 or PILOTDS3), Data Set 4 (NSSIDS4 or PILOTDS4), and a data set of
field observations recorded by field crews about the immediate watershed area and sample site sub-
strate (NSSIFSO). The decision as to which data set to use depends on the intended use. To
examine population attributes (e.g., means, percentiles, or variances), Data Set 4 should be used. It is
important that the statistical weighting factor (W) be used when examining population characteristics.
Estimating population parameters without using sample weights can lead to biased estimates and
inaccurate interpretation. However, if the intended use of the data is to examine the chemistry of
individual water samples, Data Set 3 may be more useful, as it contains flags and tags that identify
anomalies in the data or in the methodology used (e.g., holding time violations or analytical equipment
discrepancies).
2
A total of 225 values (0.5% of the database) across 9 variables (extractable aluminum, DOC, iron, manganese, ammonium,
nitrate, total dissolved phosphorus, and silica) had values scored to zero, with total dissolved phosphorus having the most scoring
changes, at 99 values. The bias due to this adjustment did not affect the population estimates presented. The range of negative
observations scored to zero were low in absolute value and below instrument detection limits in all cases (Cougan et al., 1988).
22
-------
3.5.1 Data Set 3 Versus Data Set 4
The information contained in Data Set 3 is verified and validated but has not undergone any type
of missing or erroneous value replacement or enhancement. Duplicate observations in Data Set 3 are
not averaged. Each observation in Data Set 3 contains data associated with a particular container of
water collected, including QC samples (e.g, blanks, duplicates, and audit samples); thus this data set
is sample or visit specific. Quality control samples, such as duplicates, can be identified using the
variable SAMCOD ("R1 routine, "D"- duplicate, "E" - suspected precipitation events). Unique obser-
vations can be identified in Data Set 3 by using the variables for batch and sample identification
codes: BATJD and SAMJD.
Data Set 4, in contrast, is visit specific, in that observations are intended to portray the chemistry
that existed at the time of a visit to a sample reach. Quality control samples, except for field duplicate
samples, have been removed in Data Set 4. In addition, field duplicate (QC) samples are averaged
with the corresponding routine sample, erroneous and missing values are replaced, and impossible
negative values are scored to zero. Because duplicate samples were averaged with routine samples,
batch and sample identification codes are not included in Data Set 4. As in Data Set 3, data for
multiple visits are retained (not averaged). Observations are uniquely identified using the stream
identification code (STRMJD) and the sample visit number (SAMRN).
3.5.2 The NSS-I Pilot Survey: The Southern Blue Ridge Province
In contrast to the NSS-I, which sampled both the upstream and downstream ends of a reach, the
Pilot Survey sampled the upstream ends of only 20 of the 54 reaches during a spring index baseflow
period in 1985 (Messer et al., 1986, 1988). All 54 reaches, however, were sampled at both the
upstream and downstream points during the summer 1985 (versus spring) index period. To provide
population estimates compatible with the spring upstream estimates for other subregions, empirical
relationships were used to synthesize missing values for 22 chemical variables for the 34 upstream
sites not sampled during the spring index period. These synthesized values are not included in the
enhanced database, but are provided as a separate small data set. The calculation of this
synthesized data set is detailed in Appendix B of Kaufmann et al. (1988). The results presented for
the Southern Blue Ridge in the NSS final report were generated using a data set made by appending
SBRSYN to NSSIDS4.
It is important to note that after the Pilot Survey report was published, a number of data values
were updated, resulting in a change in the estimated total number of reaches in the Southern Blue
Ridge Province from 2,021 in the Pilot Survey report to 2,031 in the NSS-I report. These updated
values were incorporated into data used in the NSS-I report to describe the Southern Blue Ridge Prov-
ince (subregion 2As). The principal updates included the revising of one direct drainage value, a1 (A1
23
-------
in the database), and a refinement in the method used to calculate ANC and BNC from Gran trtration
data. The differences between the data sets used for the NSS-I report and the Pilot Survey report are
discussed in Appendix C. Because ax is used in calculating the final sample weight of each reach
(the number of reaches the observation is estimated to represent in the target population), the total
number of estimated reaches in the Southern Blue Ridge Province changed by 10 reaches. The Pilot
survey data sets, PILOTDS3 and PILOTDS4, were used in generating results presented in the Pilot
Survey report (Messer et al., 1986). The revised data for the Southern Blue Ridge are included only in
NSSIDS4. This is the data set used in making estimates presented in the NSS-I report (Kaufmann et
al., 1988).
24
-------
SECTION 4
DEFINING THE TARGET POPULATION
4.1 EVALUATION PROCESS
This section discusses how the Stage II probability sample of reaches was evaluated and refined
to a subset that represents the NSS-I target population of stream reaches. Although many nontarget
reaches were screened out (using map information) during the Stage I selection process, it was
necessary to further refine the set of "selected" reaches, based on field and chemical data, to best
represent the target population of interest. In addition to identifying erroneous data, the validation
process helped to identify unusual sites in the context of subregion populations. For example,
reaches impacted from acid mine drainage comprise a portion of the total population for which the
additional impact from acidic deposition is extremely difficult to ascertain. Therefore, these reaches
were not included in estimates of the target population. Chemical data for such sites were not deleted
from the NSS-I database, but were marked as noninterest to allow their inclusion or exclusion in
statistical analyses with variables DRPCDE and SIT_CLS.
4.2 IDENTIFYING NONINTEREST OBSERVATIONS AND SITES
Noninterest observations were identified based on chemistry, field observations, and the
investigation of watershed characteristics. The following basic criteria were used to identify sites and
observations not included in estimates of the NSS-I target population:
Intermittent: reaches that were at least 90% dry or stagnant.
• High specific conductance: reaches having an in situ specific conductivity > 500 \iS cm"1
(e.g., reaches contaminated by oil well brine, industrial pollution, severe acid mine drainage).
• Episode (other than spring baseflow conditions): reach chemistry influenced by a precipita-
tion event at or near the time of sampling (e.g., high turbidity and high flow) (Section 4.2.1).
• Low pH: reaches having a field pH value of 3.3 or less [e.g., severe acid mine drainage
impact (Section 4.2.2)].
• Tidal influence: coastal reaches with water chemistry influenced by seawater (e.g., specific
conductivity greater than 250 ^S cm"1).
• Reservoir: reach inundated by water project.
No channel, dry: no sample could be collected because of a lack of water (dry) or a lack of
explicit stream channel with flowing water (swamp).
25
-------
4.2.1 Episode Identification
Because the NSS-I target population estimates were intended to represent spring baseflow
chemistry, field sampling was explicitly avoided during precipitation episodes. Although great care
was taken not to collect samples during a precipitation episode, such conditions might not have been
apparent to a sampling team. Samples inadvertently collected during a precipitation or snowmelt
episode, or influenced by one, were identified as those meeting ajl of the following criteria:
• Site identified as a validation outlier (see Section 3.2).
• Change in stream gauge height of 7.5 cm or more between site visits, supporting evidence
of flood stage, except for the Southeast sites where only one visit to a site was made.
• Field comments indicating a precipitation or snow melt event within 24 hours of sampling.
• Unusually high concentrations of turbidity, total aluminum, manganese, or iron relative to
other visits at the same site or its corresponding upstream or downstream node.
4.2.2 Acid Mine Drainage
Streams impacted by acid mine drainage (AMD) (i.e., those distinguished by low alkalinity and pH
and markedly high sulfate) comprise a category that had to be identified in order to distinguish them
from streams impacted by acidic deposition. Sites impacted by AMD were excluded in the calculation
of NSS-I population estimates. These streams were identified as those meeting aN of the following
criteria, the rationale for which is discussed in subsection 9.3.1 of Kaufmann et al. (1988):
• ANC < 0 jieq L'1
• Sum of base cations > 400 ^eq L"1
• Sulfate/sum of anions > 75%
DOC < 5 mg L"1
[S042~] > 300 [ieq L'1 in the Mid-Atlantic; [SO42-] > 200 jieq L'1 in the Southeast
• Mining activity confirmed by means of maps, field visits, or aerial photographs
4.3 CLASSIFICATION OF NONINTEREST OBSERVATIONS IN THE NSS-I DATABASE
Once specific site exclusion criteria were evaluated, observations classified as noninterest were
assigned a site class code (SIT_CLS) to identify specific noninterest conditions. The first part of this
code identifies the noninterest features and the second, the reach end or node (upstream or down-
stream) at which the condition was noted. These features are coded in the NSS-I database as:
26
-------
Code Condition
A Sample point impacted by acid mine drainage
C High conductivity at sample point (> 500 \iS cm"1)
I Intermittent flow
T Site chemically impacted from tidal activity
R Random miss (see discussion below)
S Special cases requiring the removal of specific observation (e.g., visit made to
the wrong reach)
It is important to note that a site was considered sampled if a field crew visited a mapped
sampling point and obtained information about the site, regardless of the presence of water. There
were, however, four cases in which field crews were not able to visit the mapped sampling point (e.g.,
reach not found by field crews) and thus, no information about these missed sites was obtained.
Such occurrences were considered to occur at random and each one is identified as a random miss
marked with an 'R' in the SIT_CLS variable.3
The second character of the site class code designates the point on the reach at which the
noninterest condition existed: "1" for the upstream site, "2" for the downstream site, or "3" for both
sites. For example, a reach assessed as having high conductivity might have a SIT_CLS code of C1
to note the condition at the upstream point, C2 to note the condition at the downstream point, or C3
to note the condition at both points. When multiple noninterest conditions occurred on a reach,
these codes were combined, (e.g., high conductivities found at both sample sites, in addition to acid
mine drainage evidence at a downstream end would be classified as "A2C3").
The second character in SIT_CLS provides exclusion criteria for the entire reach regardless of the
observation (or node) being examined. In the above example, the SIT_CLS code "A2C3" would
appear in all observations for that reach.
n~he occurrence of a random miss within a stratum required an adjustment in the final weighting factor for all reaches in that
stratum (Overton, 1987).
27
-------
4.4 DROP CODE VALUES
The drop code variable, DRPCDE, was created to allow the selective exclusion of noninterest
observations in statistical analyses. All observations were assigned a code to indicate whether they
are to be considered part of the target population or the noninterest group of reaches.
Population
Target
Noninterest
Noninterest
Noninterest
Drop
Code
0
1
2
3
Noninterest 4
Noninterest 5
Noninterest 13
Exclusion Criteria Description
NSS-I target observation for both the upper and lower node.
Alternate node of noninterest site (see discussion that follows).
Noninterest sites (Kaufmann et al., 1988, Section 3.2).
Sites acidic due to acid mine drainage (Kaufmann et al., 1988, Subsection
9.3.1).
Pilot Survey nonspring index data (water samples not collected between
March 15 and May 1) (in Kaufmann et al., 1988, Subsection 1.3.2.1).
NSS-I special interest sites (Kaufmann et al., 1988, Section 2.6).
Combination of DRPCDE 1 and DRPCDE 3 (i.e., an alternate node of a
noninterest site is acidic due to acid mine drainage).
The set of observations considered to represent the target population of streams are those with a
DRPCDE value of "1" or less. Population estimates were made for upstream and downstream
locations as separate populations. However, when conducting analyses that examine those reaches
at which both the upstream and the downstream ends fit the target criteria, the data can be subsetted
by including only those observations with a DRPCDE value of 0.
Because specific noninterest conditions (e.g., intermittent flow) may eliminate either reach end,
upstream or downstream, as a target, it is possible to have one end of a reach included in the target
population, but not the other. For example, estimates were made in the NSS-I for different numbers of
upstream and downstream ends.
28
-------
SECTION 5
DATABASE APPLICATION
5.1 GENERATING THE DATA SET OF INDEX VALUES FROM THE REACHES SAMPLED
This section discusses the method used to calculate NSS population estimates. An important
aspect of the NSS data design is that population estimates are based on a single observation for each
stream reach. In the NSS-I report, separate estimates are made based on observations representing
upstream and downstream sample sites. Using Data Set 4 (NSSIDS4, SBRSYN, PILOTDS4), this is
accomplished by averaging multiple observations for each sample site. The resulting data set will
contain information for the upstream and downstream ends of reaches, which are treated as repre-
senting separate populations of interest. We chose not to average upstream observations with
downstream observations, but doing so does not violate the NSS design.
The process of generating an "indexed" data set includes:
• Identifying appropriate data sets (NSSIDS4 and SBRSYN for NSS-I estimates and PILOTIDS4
for NSS-I Pilot Survey estimates).
Subsetting observations with a DRPCDE < 1 (or deleting observations with DRPCDE > 2).
• Calculating a mean value of the parameter(s) of interest (e.g., ANC, pH) for each STRMJD.
Each resulting observation should contain a value for the parameter(s) of interest (e.g., ANC,
pH), as well as data for SUBJD, STRMJD, NODE, W, STRATUM, and any other classification
unit desired (e.g., STATE1, RCH_HW). Note: It is recommended that the mean of multiple
pH values be calculated as the mean of hydrogen ion concentrations and converted into pH.
Table 5-1 lists the number of visits, total observations, high-interest samples, and nontarget
samples, and the number of indexed observations by subregion.
5.2 EXTRAPOLATION TO THE TARGET POPULATION
The following section briefly discusses the methods used for making NSS-I population estimates.
The computer programs used to generate the statistical estimates and graphics (cumulative
distribution function curves, trilinear plots, etc.) are included in Sale et al., 1990, Data Management and
Analysis Procedures.
5.2.1 Sample Weightings For Population Estimates
The NSS-I statistical design uses the attributes (e.g., chemistry, geographic features) of the 500
sampled reaches to describe the characteristics of an estimated 64,260 reaches in the target popula-
tion. In the NSS-I database, each sample reach is assigned a calculated weight that indicates how
29
-------
TABLE 5-1. NUMBER OF VISITS, TOTAL OBSERVATIONS, TARGET AND NONTARGET
SAMPLES, AND INDEXED OBSERVATIONS BY SUBREGION
8
Region
Mid-Atlantic
•
•
•
Southeast
•
N
•
N
Special Interest
Sites
Subregion
1D
2Bn
2Cn
3B
2D
2X
3A
3C
2As
SI
#of
Sample
Visits per
Site3
2
2
2
2
1
1
1
1
1-5C
1-4
#0f
Observations
Upper Lower
122
103
145
119
50
50
50
50
109d
2
122
102
144
114
50
50
50
50
233
84
#0f
Noninterest
Observations
Upper Lower
6
16
14
7
1
11
3
19
55
-
10
11
25
6
2
10
3
16
78
-
# of Target
Observations
Upper Lower
116
87
131
112
49
39
47
31
54d
-
112
91
119
108
48
40
47
34
155
-
#of
Spring Indexed
Observations'3
Upper Lower
58
44
67
57
49
39
47
31
54d
--
56
47
61
58
48
40
47
34
54
-
a The number of sample visits correspond to the variable SAMRN in the NSS-I and Pilot Survey data base.
b The number of observations in the "indexed" (averaged) data set.
c Observations for 2As, the Southern Blue Ridge Province, are identified in the data base as 0-4. Only sample visits 1-3 are within the spring
index period.
d 34 observations for the Southern Blue Ridge were synthesized for the NSS-I data report (see Section A.1).
-------
many other reaches in the population are represented by that reach. This weighting factor is inversely
proportional to a reach's probability of being selected and is the product of sample weights for two
stages of sampling. Recall that in the first stage, the probability of selecting a reach was directly pro-
portional to the direct drainage area of the reach, a/. In the second stage, sites were selected with
probabilities inversely related to their first stage inclusion probability, largely equalizing the final
sample weightings. After completion of the Stage I selection process, a few a1 values were updated.
As in the Pilot Survey, the final sample weight of a reach is directly related to its a1. To adjust for
these changes, an update was made to the Stage II conditional weights. Details of estimation and the
statistical foundation are provided in Kaufmann et al. (1988), Section 2, Overton (1986, 1987), Blick et
al. (1987), Overton and Stehman (1987), and Stehman and Overton (1987a, 1987b). Two weighting
variables are present in the database, "WC", the conditional Stage II subsampling weight, and "W".
The final NSS-I sample weight, W, is used in extrapolating from the sample to the population:
W = WC (64/a^
where:
64 = Area (mi2)/grid point
Note: If a1 < 0.2 mi2, a1 was scored to 0.2 mi2 in the calculation of W. Although not present
in the NSS database, an unscored weight can be generated by multiplying WC by
64/ar
Whenever statistical descriptions of NSS-I reach populations are made, the sample weight, W,
should be used. It is recommended that weights be used when calculating frequency distributions
and other univariate analyses of chemical data, including the calculation of means. However, NSS-I
data users are cautioned to assess the use of weights in bivariate or multivariate explorations of the
data. Scatter plots, correlations, and regressions are affected by the use of weights and careful
thought should be given, especially if data are combined over strata (e.g., across subregions). Keep
in mind that unweighted parameter estimates do not necessarily represent those of the population.
An important issue in the use of weights to generate population statistics is how to calculate
standard errors associated with estimates. The NSS-I bases standard error estimates on a variation of
the Horvitz-Thompson variance estimator (Overton, 1987). Users are cautioned in the use of sample
weights in statistical analyses (see Section A.4). Any analyses that incorporate a standard error
estimate must use the standard error algorithm used in the NSS.
In a small number of reaches, a1 was less than 0.2 mi. In these cases the value 0.2 mi was substituted as the divisor to
reduce the variance of the population estimates (Overton, 1987, discusses the rationale and application of this scoring process)'
31
-------
5.2.2 Estimating the Target Population
In the NSS-I, the basic parameter estimate (and variance) of the population of interest is based
on the total number of reaches, as portrayed by the upstream sites or downstream sites separately, or
the total of any attribute of interest, using the Horvitz-Thompson (1952) estimator:
Ty = I
ieS
where:
ty is the estimate total of any attribute, y, over the population.
y is any attribute of interest over the sample, S. When making estimates of the number of
reaches, y = 1; for estimates of reach length, y = RCH_LN; for total watershed area
estimates, y = A_WS; for direct drainage area, Y = A1.
w is the sample weight assigned to each reach.
Z indicates summation over the entire sample (or subset) of target reaches.
A primary product of the NSS-I, these estimates of the distribution of target population attributes
were made by assigning different definitions to y (e.g., total length), and summing over different sets
of sample units, S (e.g., Subregion). The NSS-I examines several attributes of the target population:
• Number of target reaches N(x), y = 1: N(x) = Z Wj
X
-------
X < x refers to those conditions in which the estimate is incrementally made over a
range of an attribute (i.e., N(x) is the estimated number of reaches in the target
population with a value < x).
• X\i is the topographic watershed drainage above a sample site ('A_WS" in the NSS
database).
/ is the length of reach segment between mapped confluences ('RCHJJSI1 in the
NSS-I database).
The distribution of attributes across a range of chemistry is calculated by computing estimates
for each value of x, and then, in a cumulative manner, dividing by the total number of estimated
reaches (or sum of weights). The following shows a distribution for example data.
ANC
-50.0
-25.5
0.0
25.7
50.3
200.1
Sample
Weight
15
150
50
250
35
215
Cumulative
Weight
15
165
215
465
500
715
Percent
of Total
0.021
0.209
0.069
0.348
0.049
0.301
Cumulative
Percent
0.021
0.231
0.301
0.650
0.699
1.000
For each distribution, it is possible to estimate quartiles, the median, and quintiles referred to as
01, Q2, Q3, Q4. The median is the value of x such that it is 1/2 or 0.5 on a cumulative proportion F(X)
curve (Section 5.3.1). These statistics are defined and presented for all distributions in Sale et al.
(1988). Additionally, the mean and standard deviation of the variable x in the population can be
estimated.
Mean(x)
SD(x) =
5.2.3 Variance Estimates and Confidence Bounds
The variance estimates, leading to estimated standard errors (SE) were obtained by application
of an original variation of the Horvitz-Thompson variance estimator (Overton, 1986; Stehman and
Overton, 1987a). The formula for the estimated variance is:'
V(Ty) = Z yi2wi(wi-1)+ Z ZyiyjVij
ieS ieSjeS
33
-------
where:
if observations i and j are from the same stratum (Subregion and ANC), then
v.. = [(Wj+Wj)/2 - WjWj]/(n-1), else, if i and j are from different strata then v^ = 0.
n is the Stage II effective sample size for each stratum (from Table 2-1).
Simply put, the first portion of this equation is a variance component while the second is a covariance
component, specific to the observations of each stratum.
Once variance estimates are made, the associated standard errors and one-sided 95% upper
confidence bounds can be generated. The following statistics are associated with N, the estimated
total number of target reaches, estimated as [t , where y = 1 (see Section 5.2.2)]:
• The weighted cumulative proportion estimates calculated as F (x) = N(x)/N.
• The standard error for this estimate (A (x)n), calculated as SE(N(x)) = vV(N(x)).
• A one-sided 95% upper confidence bound5, calculated (assuming a normal distribution) as
Nu(x) = N(x) + 1.645[SE(N(x))].
• Weighted percentile estimates of one-sided 95% upper confidence bound, calculated as
Fu(x) = Nu(x)/N.
Although not presented in the NSS-I report, a two-sided 95% confidence bound can be generated as
NL(X) = N(x) ± 1.96[SE(N(x))].
5.3 DESCRIBING THE TARGET POPULATION
This section outlines ways the NSS-I data can be used to describe the target population of
streams. The database structure facilitates the examination of the target population in a number of
ways. Whereas the range of target streams is explicitly defined in terms of size, watershed area, and
general water quality conditions, the Survey design allows the end user to examine the entire popu-
lation of sampled streams, and not just those used to calculate refined target estimates. For example,
if one is interested in examining that portion of the NSS-I population impacted by acid mine drainage,
For the majority of the NSS-I estimates, only a one-sided upper 95% confidence bound was generated as depicted in distribution
plots in Sale et al. (1988). A lower one-sided 95% confidence bound can be generated, if needed, by assuming that a lower
confidence bound is approximately equal to N -1.645 [SE(N)].
34
-------
the data can be subsetted using the appropriate drop codes and weights. Examples of other NSS-I
populations might include:
• The initial NSS-I target population (i.e., the population represented by the complete set of
sampled Stage II reaches, including noninterest and interest sites).
[Subsetting observations based on a DRPCDE < 4 and then indexing]
• The set of headwater reaches.
[Subsetting observations based on RCH_HW = 1 and then indexing]
• The set of reaches located within a specific geographic area (e.g., NSS-I reaches whose
upstream or downstream ends are located in Maryland).
[Subsetting observations based on STATE1 = 'MD']
The set of reaches defined by some chemical attribute (i.e., reaches with DOC < 10 mg L"1).
[Subsetting observations based on DOC11 < 10]
Again, it is important to remember that during examination of the population represented by the
sampled reaches, the sample weighting factor, W (contained in Data Set 4), must be used.
Table 5-2 presents estimates of the target stream population by subregion and node for the total
number of sites and total reach length. Estimates were made using the equations presented in
Section 5.2. The set of streams represented in Table 5-2 is identified in the NSS-I database as those
sites with a DRPCDE value < 1. Special interest sites are not used in making statistical estimates of
the target population. This data subset includes only those sites that represent spring baseflow con-
ditions and are not grossly impacted from influences (e.g., acid mine drainage, tidal influence) that
mask the effects of acidic deposition.
5.3.1 The Cumulative Distribution Function (CDF) Curve
In the NSS-I, distributions are generated for measured attributes using the general estimators
(e.g., numbers, length, watershed area) and presented using the cumulative distribution function
curve, F(x), sometimes called a cumulative proportion (Overton, 1989). The distribution, F(x), shown in
Figure 5-1, is interpreted as the proportion of target reaches in the population having the attribute X <
x. To read this figure, pick a value of x of the attribute X, along the horizontal axis (ANC in this
example) and read the y-axis value of the two curves, F(x) at this value. The F(x) is the estimated
proportion of reaches in the population with a value of the attribute equal to or less than X. In this
example, the median or 50th percentile (i.e., F(x) = 0.5) is read as 257 jieq L"1. The estimated
number of downstream reach ends, which is less than or equal to this value, or N(x), can be
35
-------
TABLE 5-2. TOTAL RESOURCE ESTIMATES OF THE NSS REFINED TARGET POPULATION
Subregion
1D
2Bn
2Cn
3B
2As
2D
2X
3A
3C
au
n =
/•V
N
SE(N) =
ys.
SE(L) =
Node8 n
U 58
L 56
U 44
L 47
U 67
L 61
U 57
L 58
U 54
L 54
U 49
L 48
U 39
L 40
U 47
L 47
U 34
L 31
N
3244
3235
13038
13992
8663
8488
11284
11287
2031
2031
4204
4116
4936
5057
7515
7515
1727
1555
upstream or "upper node" sample sites; L
number of sampled stream
estimated target population
standard error of N.
reaches.
size.
estimated total stream reach length of the
(reach length).
standard error of L.
SE(N)
347
347
1249
1213
807
814
1078
1078
326
326
406
410
529
526
650
650
437
306
L
15270
15144
32687
36405
22373
21738
40296
40344
9036
9036
22753
22480
21892
23015
33531
33531
4312
4820
= downstream or "lower
target population, based
SE(L)
1911
1912
4492
4678
2732
2746
5799
5799
960
960
2485
2507
2807
2895
4402
4402
690
731
node" sample sites.
on the variable RCH_LN
36
-------
1.0"
0.9-
0.8
0.7
o
'
'•gO.6
o.
o 0.5
O
0.4
0.3-
0.2
0.1
o.o-
Downstream Population
Number Of Reaches
Proportion £. X
Upper 95% CL
-100 0 100 200 300 400 500 600 700 800 900 1000
ANC (Meq L"1)
A cumulative distribution function curve can be calculated for the NSS target population of downstream sites based on the
following.
1. Subset the data such that:
* Only those observations that have a DRPCDE <, 1 are included, and
* Only those observations for the downstream points (NODE=L) of interest.
2. Generate a data set of mean chemistry values of interest for each reach (by the variable STRMJD in the database).
Note: be sure to maintain weights in the new data set. At this point, only one observation should exist for each
downstream sample site.
3. Identify the population subset to analyze (e.g., streams in Maryland).
4. Sort the observations based on the parameter of interest (i.e., ANC).
5. In a cumulative manner by each observation of x, divide the weight by the estimated population total (i.e., a
cumulative percentage is calculated by summing up the weights over the distribution of interest and dividing by the
population total) such that for each value of x, F(x) = N (x)/N.
6. The resulting data can be plotted as an XY plot with the cumulative percentage (from 0 to 1) on the Y-axis and the
parameter of interest on the X-axis. Assuming that values lying between the calculated percentages are located
along a straight line between plotted points, these points are connected by a straight line.
7. The upper one-sided confidence bound is estimated for each value of x (Section 5.2.3) and a cumulative proportion
is estimated [on each value of N(x)], based on the estimate of the total number of sites, N.
Figure 5-1. Calculation of an example cumulative distribution function curve for downstream sites.
37
-------
calculated by multiplying F (x) by Ntota). The dotted line in this figure can be used to estimate a one-
sided 95% confidence bound on N(x) representing Nu(x). In this example, Nu (x) for the median is
calculated as 0.76 * Ntota).
In addition to distributions presented as F(x), some are presented as an inverse, or descending,
cumulative proportion, or 1 - F(x). For these, read the distribution as the estimated proportion of
reaches having values equal to or greater than x (X > x).
5.3.2 Length Estimates
In addition to distributions in terms of the numbers and percentages of stream reaches, estimates
were made of the combined length of reaches in the target population (Overton, 1989). These length
estimates were calculated using two different approaches that yielded different estimates.
5.3.2.1 First Approach: Length Estimates Based on Node Chemistry-
The first approach used to calculate the length estimates presented in Sale et al. (1988), assigns
a measured chemical parameter value (e.g., ANC) at the downstream node to the entire length of
sample reach (using the variable RCH_LN).
1. Subset over unit of sample interest (e.g., subregion 2As) for observations with DRPCDE < 1.
2. Generate data set of mean values (Section 5.1) with a single value for each upstream and
downstream end.
3. From data set of index values, generate data set that merges the upstream and downstream
data into a single observation, keeping the chemistry associated with each node identified
separately. For each observation, there will be data from upstream and downstream ends.
4. Calculate weighted length for each reach by multiplying each value of RCH_LN by reach
sample weight, W.
5. Sort data generated in step 4 by the chemistry value of interest and generate a CDF (using
method described in Section 5.3.1).
5.3.2.2 Second Approach: Length Estimates Based on Interpolated Length-
The second method (Illustrated in Kaufmann et al., 1988) assumes the chemical concentrations
change in a linear fashion from the upstream sampling point to the lower one. This method estimates
a series of linearly interpolated chemical values along a reach length between the upstream and
downstream ends. Each segment is then assigned the reach's respective sample weight, thereby
extrapolating to the target population of reaches. Figure 5-2 presents cumulative length estimates for
38
-------
Interpolated Length Estimates
SUBREGION=2AS
O
|
m
O
1.0-
0.9-
0.8
0.7-
0.6-
0.5-
0.4
0.3-
0.2-
0.1-
0.0 -,
Upper Node Based Estimate
Interpolated Estimate
Lower Node Based Estimate
50
100
ANC
150
L"1)
200
250
300
Figure 5-2. Interpolated length estimates for subregion 2As. Overlay of three CDFs of ANC con-
centration versus cumulative reach length based on (a) chemistry from upstream sites
or nodes, (b) chemistry from downstream sites or nodes, and (c) an interpolated
chemistry and length value between the upstream and downstream site or node
values.
39
-------
subregion 2As, based on lower node chemistry, upper node chemistry, and interpolated length
chemistry. As expected, the interpolated length distributions generally are bounded by the upstream
and downstream distributions. Note: when distribution estimates are made for pH, the interpolation is
actually done on [H+] and converted to pH values (based on PHSTVL). The following two methods
can be used to estimate interpolated NSS-I reach lengths.
Estimated Interpolated Length: Method 1 - Interpolation of Reach Segments
1. Subset over unit of sample interest (e.g., Subregion 2As) for observations with DRPCDE < 1.
(Sites with DRPCDE = 0 have both upstream and downstream reach ends included in the
target population.)
2. Generate data set of mean values (Section 5.1) with a single value for each upstream and
downstream end.
3. From data set of index values, generate data set that merges the upstream and downstream
data into a single observation, keeping the chemistry associated with each node identified
separately. For each observation, there will be data from upstream and downstream ends.
4. Divide each reach length (using variable RCH_LN) into 0.1 km segments. Determine the
difference for the chemistry value of interest between upstream and downstream ends for
each reach and divide by number of 0.1 km segments estimated for reach. Resulting value
represents estimated incremental change that occurs over each 0.1 km segment. Starting at
one end, assign a chemistry for each segment, incrementally estimating each segment
chemistry with each increment in length.
5. Sort data set generated in step 4 by the chemistry value and generate a CDF (using the
method described in Section 5.3.1).
Estimated Interpolated Length: Method 2 - Simple Linear Interpolation
1. Subset over unit of sample interest (e.g., Subregion 2As) for observations with DRPCDE < 1.
2. Generate data set of mean values (Section 5.1) with a single value for each upstream and
downstream end.
3. From data set of index values, generate data set that merges upstream and downstream
data into a single observation, keeping the chemistry associated with each node identified
separately. For each observation, there will be data from upstream and downstream ends.
4. Using reach length (RCH_LN) and difference in chemistry between ends of each reach, it is
possible to interpolate a length of each stream that is < a particular reference value. By
doing this over an incremented range of chemistry, interpolated length estimates can be
made for an entire subset of data. (In Figure 5-2, estimates were made for each value of
ANC from 0 to 1,000, in 1 ^eq ^increments.) Note: this can only be done if the range of
reference values encompasses the range of observed values in the data. Also, the resolution
of this method is limited by the increment size of the reference values (in this example, 1
L'1.
40
-------
5. Sort data set generated in step 4 by the chemistry value and generate a CDF (using the
method described in Section 5.3.1). Note: a standard error can be calculated for
interpolated length estimates by assigning interpolated length as the attribute of interest (see
Section 5.2.3).
41
-------
NSS DATABASE GUIDE REFERENCES
American Society for Testing and Materials. 1984. Standard specification of reagent water. D1193-77
(reapproved 1983). In: Annual Book of ASTM Standards. Vol. 11.01. Philadelphia,
Pennsylvania.
Arent, L.J., M.D. Morison, and C.S. Soong. 1988. Eastern Lake Survey - Phase II, National Stream
Survey - Phase I: Processing Laboratory Report. EPA/600/4-88/025. U.S. Environmental
Protection Agency, Washington, D.C. 86 pp.
Binder, D., J. Kovar, S. Kuar, D. Paton, and A. van Baaren. 1987. Analytic uses of survey data: a
review. Pages 243-264. In: I.B. MacNeill and G.J. Umphrey, eds. Applied Probability, Stochastic
Processes, and Sampling Theory. D. Reidel Publishing Co.
Blick, D.J., J.J. Messer, D.H. Landers, and W.S. Overton. 1987. Statistical basis for the design and
interpretation of the National Surface Water Survey, Phase I: Lakes and Streams. Lake Reserv.
Manage. 3:470-475.
Cougan, K.A., D.W. Sutton, D.V. Peck, V.J. Miller, and J.E. Pollard. 1988. National Stream Survey -
Phase I: Quality Assurance Report. EPA/600/4-88/018. U.S. Environmental Protection Agency,
Washington, D.C. 205 pp.
Cummins, K.W. 1962. An evaluation of some techniques for the collection and analysis of benthic
samples with special emphasis on lotic waters. Am. Midi. Nat. 67:477-504.
Drouse, S.K., D.C. Hillman, LW. Creelman, and S.J. Simon. 1986. National Surface Water Survey:
National Stream Survey (Pilot, Middle-Atlantic Phase I, Southeast Screening, and Episodes Pilot).
Quality Assurance Plan. EPA/600/4-86/044. U.S. Environmental Protection Agency, Washington,
D.C. 215pp.
Drouse, S.K. 1987. Evaluation of Quality Assurance and Quality Control Sample Data for the National
Stream Survey (Phase I - Pilot Survey). EPA/600/8-87/057. U.S. Environmental Protection
Agency, Washington, D.C. 45 pp.
Hagley, C.A., C.L Mayer, and R. Hoenicke. 1988. National Surface Water Survey: National Stream
Survey (Phase I, Southeast Screening, and Episodes Pilot). Field Operations Report. EPA/600/4-
88/023. U.S. Environmental Protection Agency, Washington, D.C. 38 pp.
Hillman, D.C., S.L Pia, and S.J. Simon. 1987. National Surface Water Survey: National Stream
Survey (Pilot, Middle-Atlantic Phase I, Southeast Screening, and Middle-Atlantic Episodes Pilot).
Analytical Methods Manual. EPA/600/8-87/005. U.S. Environmental Protection Agency,
Washington, D.C. 265 pp.
Kanciruk, P., R.J. Olson, and R.A. McCord. 1986. Quality control in research databases: the U.S.
Environmental Protectional Agency National Surface Water survey experience. In: W.K. Michener
(ed). Research Data Management in the Ecological Sciences, University of South Carolina Press,
Columbia, South Carolina pp. 193-207.
Kaufmann, P.R., AT. 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.
42
-------
Knapp, C.M., C.L Mayer, D.V. Peck, J.R. Baker, and G.J. Filbin. 1987. National Surface Water Survey:
National Stream Survey (Pilot Survey). Field Operations Report. EPA/600/8-87/019. U.S.
Environmental Protection Agency, Washington, D.C. 66 pp.
Kramer, J.R. 1984. Modified Gran analysis for acid and base titrations. Environmental Geochemistry
Report No. 1984-2. McMaster University, Hamilton, Ontario, Canada.
Landers, D.H., 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. 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.
176pp.
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.
Messer, J.J., and K.N. Eshleman. 1987. The Feasibility of Quantifying the Regional Extent, Magnitude,
Duration, and Frequency of Episodes: Results of the National Surface Water Survey Pilot Studies.
Presentation at NAPAP Aquatic Effects Task Group VI Peer Review, May 17-23, New Orleans,
Louisiana.
Messer, J.J., C.W. Ariss, J.R. Baker, S.K. Drouse, K.N. Eshleman, P.R. Kaufmann, R.A. Linthurst, J.M.
Omernik, W.S. Overton, M.J. Sale, R.D. Schonbrod, S.M. Stambaugh, and J.R. Tuschall, Jr. 1986.
National Stream Survey Phase I - Pilot Survey. EPA/600/4-86/026. U.S. Environmental Protection
Agency, Washington, D.C. 179 pp.
Messer, J.J., D.H. Landers, R.A. Linthurst, and W.S. Overton. 1987. Critical design and interpretive
aspects of the National Surface Water Survey. Lake Reserv. Manage. 3:463-469.
Messer, J.J., C.W. Ariss, J.R. Baker, S.K. Drouse, K.N. Eshleman, A.J. Kinney, W.S. Overton, M.J. Sale,
R.D. Schonbrod. 1988. Stream chemistry in the Southern Blue Ridge: feasibility of a regional
synoptic sampling approach. Water Resour. Bull. 24:821 -829.
Nathan, G. 1988. Inference based on data from complex sample. Pages 247-266. In: P.R.
Krishnaiah and C.R. Rao, eds. Designs Hand of Statistics. Volume 6. Elseveier Science
Publishers B.V.
O'Dell, J.W., J.D. Raff, M.E. Gales, and G.D. McKee. 1984. Technical Addition to Methods for the
Chemical Analysis of Water and Wastes. Method 300.0. The Determination of Inorganic Anions
in Water by Ion Chromatography. EPA/600/4-85/017. U.S. Environmental Protection Agency,
Cincinnati, Ohio.
Oliver, B.G., E.M. Thurman, and R.K. Malcolm. 1983. The contribution of humic substances to the
acidity of colored natural waters. Geochim. Cosmochim. Acta 47:2031-2035.
Omernik, J.M., and A.J. Kinney. 1985. Total alkalinity of surface waters: a map of the New England
and New York Region. EPA/600/D-84/216. U.S. EPA Environmental Research Laboratory,
Corvallis, Oregon.
Omernik, J.M., and C.F. Powers. 1983. Total alkalinity of surface waters»a national map. Ann. Assoc.
Am. Geog. 73:133-136.
43
-------
Overton, W.S. 1986. A Sampling Plan for Streams in the National Stream Survey. Technical Report
114. Department of Statistics, Oregon State University, Corvallis, Oregon. 18 pp.
Overton, W.S. 1987. A Sampling and Analysis Plan for Streams in the National Surface Water Survey.
Technical Report 117. Department of Statistics, Oregon State University, Corvallis, Oregon. 50
PP-
Overton, W.S., and S.V. Stehman. 1987. An Empirical Investigation of Sampling and Other Errors in
the National Stream Survey; Analysis of a Replicated Sample of Streams. Technical Report 119.
Department of Statistics, Oregon State University, Corvallis, Oregon. 21 pp.
Overton, W.S. 1989. Effects of Measurement and Other Extraneous Errors on Estimated Distribution
Functions in the National Surface Water Surveys. Technical Report 129. Department of Statistics,
Oregon State University, Corvallis, Oregon. 21 pp.
Pollack, A.K., and S.C. Grosser. 1988. National Stream Survey Data Base Audit. SYS APP-88/088.
Systems Applications, Inc., San Rafael, California. 49 pp.
Sale, JJ. (ed.). 1990. Data Management and Analysis Procedures for the National Stream Survey.
Oak Ridge National laboratory, Oak Ridge, TN. (Draft)
Sale, M.J., P.R. Kaufmann, H.I. Jager, J.M. Coe, K.A. Cougan, A.J. Kinney, M.E. Mitch, and W.S.
Overton. 1988. Chemical Characteristics of Streams in the Mid-Atlantic and Southeastern United
States. Volume II: Streams Sampled, Descriptive Statistics, and Compendium of Physical and
Chemical Data. EPA/600/3-88/021 b. U.S. Environmental Protection Agency, Washington, D.C.
595 pp.
SAS Institute, Inc. 1985. SAS User's Guide: Statistics, Version 5 Edition. SAS Institute, Inc., Gary
North Carolina. 956 pp.
Shreve, R.L 1966. Statistical law of stream numbers. J. Geol. 74:17-37.
Skougstad, M.W., J. Fishman, LC. Friedman, D.E. Erdmary, and S.S. Duncan, eds. 1979. Methods
for Determination of Inorganic Substances in Water and Fluvial Sediments: Techniques of Water
Resources Investigations of the United States Geological Survey. Book 5r Chapter A1. U.S.
Government Printing Office, Washington, D.C.
Stehman, S.V. and W.S. Overton. 1987a. Estimating the variance of the Horvitz-Thompson estimator
for variable probability, systematic samples. Proceedings of the Survey Research Section. 1987
ASA Annual Meetings, San Francisco, CA.
Stehman, S.V. and W.S. Overton. 1987b. An Empirical Investigation of the Variance Estimation Meth-
odology Prescribed for the National Stream Survey: Simulated Sampling from Stream Data Sets.
Technical Report 118. Department of Statistics, Oregon State University, Corvallis, OR. 19 pp.
Strahler, A.N. 1957. Quantitative analysis of watershed geomorphology. Trans. Am. Geophys. Union
38:913-920.
U.S. Environmental Protection Agency. 1983. Methods for Chemical Analysis of Water and Wastes.
EPA/600/4-79/020. U.S. Environmental Protection Agency, Cincinnati, Ohio.
U.S. Environmental Protection Agency. 1987. Handbook of Methods for Acid Deposition Studies:
Laboratory Analyses for Surface Water Chemistry. EPA/600/4-87/026. U.S. Environmental Protec-
tion Agency, Office of Research and Development, Washington, D.C. 342 pp.
44
-------
APPENDIX A
NOTES OF CAUTION AND ISSUES OF INTEREST TO NSS DATA USERS
Working with the NSS data can be very complex. This appendix is provided to assist the data
user with issues and cautions pertaining to data use. Although some topics are discussed in other
sections of this guide or in other documents, this section details all known issues of concern and
sources of potential problems for NSS data users. These include:
Section Description
A.1 Use of Drop Codes: DRPCDE
A.2 Generating a Working Data Set
A.3 Data Set of Field Site Observations
A.4 Use of Sample Weights: W
A.5 Chemical Variables with Similar Names
A.6 Population Estimates for Geographic Subsets of Streams
A.7 Reach Length Estimates: RCH_LN versus L2
A.8 Topographic Drainage Area Measurements: av a2, a3> a4, a5, atota, (A1, A2, A3, A4, A5,
and A_WS)
A.9 Synthesized Data for the Southern Blue Ridge (Subregion 2As)
A.10 Differences Between the NSS-I and the NSS Pilot Survey
A.11 Revisions to NSS-I Pilot Survey Data
A. 12 Comparison of Parameter Units in the NSS-I and the NLS
A. 13 Revisions to a1
A. 14 NSS Database Variable Formats
A.15 Subregion Identification Codes
A. 16 Using DIG and pH in Calculated Variables
45
-------
A.1 USE OF DROP CODE VARIABLE: DRPCDE
(Detailed In Section 4.4)
The distributed NSS-I data sets contain information for all sampled streams, including those
having sampling points or conditions considered to be 'noninteresf in the context of the NSS-I target
population. These noninterest conditions indicate influence by factors that make it difficult to discern
the impact from acid deposition (e.g, acid mine drainage, tidal influences, or sample collection during
a baseflow period other than spring). Such observations were not used in generating statistical
estimates of the NSS-I target population (in terms of status and extent) but are included in the
distributed data to allow for the examination of a variety of stream subsets.
Noninterest sites can be identified by using the variable, DRPCDE, created to allow the inclusion
or exclusion of noninterest sites from specific analyses. By subsetting the data for only those sites
with DRPCDE < 1, the resulting data set includes only those observations considered to represent
spring baseflow chemistry in the target population. This data subset will exclude data from sites
found to be grossly polluted, sampled during a summer (non-spring) base flow period, or influenced
by precipitation episode conditions. In addition, data from special interest sites were not used when
generating NSS-I target population estimates and are excluded when the drop code variable is
applied.
A.2 GENERATING A WORKING DATA SET
(Detailed in Section 5.1)
All analyses and estimates made for the NSS-I target population were generated using Data Set 4
(NSSIDS4). The NSS-I data comprises information collected from one or more visits to upstream and
downstream ends of reaches. An important aspect of using the NSS-I data to generate population
estimates is that upstream ends of reaches are considered to represent a population of stream
locations different and separate from a population of downstream ends. Therefore, a working data set
is generated in which there is only one observation for each upstream reach end and one observation
for each downstream reach end. This is accomplished by first subsetting the data using the drop
code variable and then averaging multiple visits to the same sampling site, but not upstream sites with
downstream sites. Again, it is important to remember that population estimates for upstream reach
ends were generated separately from estimates of NSS-I downstream reach ends in the final NSS
report (Kaufmann et al., 1988; Sale et al., 1988).
46
-------
A.3 DATA SET OF FIELD SITE OBSERVATIONS
(Detailed in Appendix D)
Originally categorized as "Watershed Characteristics", this information, contained in NSSIFSO,
pertains only to the area in the immediate vicinity of the sampling site. Field crews recorded observa-
tions of watershed characteristics near the site but did not perform an extensive field reconnaissance
of the entire watershed area to identify disturbances that may influence reach chemistry. These data
were collected to document potential watershed influences on sample chemistry and do not identify all
potential disturbances in the entire watershed. These observations were recorded in a standardized
format for disturbances (e.g., nearby roads, housing, agriculture, industry, and logging) and substrate
composition and vegetation coverage of the immediate sampling location. Although none of this infor-
mation has undergone the level of quality assurance that was applied to the chemistry data, such
information is helpful when interpreting individual sample chemistries.
A.4 THE USE OF SAMPLE WEIGHTS: W
(Detailed in Section 5.2.1)
Generating statistical estimates of NSS-I populations can be very complex. Because streams
were sampled with varying probabilities, the weighting factor, W, should be used to obtain unbiased
estimates. In this manner, population means, totals, and proportions can be calculated in a relatively
straightforward manner. However, standard errors of these estimates are difficult to compute. Since
most statistical packages do not compute a standard error in a manner comparable to the Horvitz-
Thompson theorem, the algorithm used in the NSS (see Overton, 1988), care must be taken when
generating such estimates. The algorithm used to estimate standard error in the NSS is described in
Section 5.2.1. The use of weights in computing regressions and other more complicated statistics is
also potentially very complex and should be undertaken with care (Binder et al., 1987; Nathan, 1988).
A.5 CHEMICAL VARIABLES WITH SIMILAR NAMES
Care should be taken in identifying the most appropriate variable for each analysis. There are as
many as five different types of variables for aluminum, pH, DIG, specific conductance, and
phosphorus, as well as the flag variables. In addition, certain variables are listed with suffixes of both
"11" and "16" (e.g., CA11 and CA16). The "11" suffix refers to values presented in the units measured
at analytical laboratories, and the "16" suffix identifies variables that have been converted or
transposed into a unit considered more appropriate for reporting results. The suffix "VAL" is used to
47
-------
refer to certain parameters measured at the processing laboratory (e.g., CONVAL, DICVAL, etc.). The
data dictionary (Appendix B) should be consulted for review of specific parameters and their units.
A.6 POPULATION ESTIMATES FOR GEOGRAPHIC SUBSETS OF STREAMS
When making estimates of the target population of reaches within a geographic subunit of the
NSS (e.g., state of West Virginia or New York), note that such estimates apply only to streams that are
located within the areas actually surveyed and that fit the target reach criteria. In other words, the
NSS-I may have sampled only part of a larger geographic area of interest (e.g., two areas in the state
of Florida) and it would not be appropriate to assume that the NSS-I estimates apply to areas outside
the NSS-I subregion boundaries (e.g., the entire state of Florida).
A.7 REACH LENGTH ESTIMATES: RCHJ.N VERSUS L2
(Detailed in Section 5.3.2)
Two different variables designate measurements of reach length in the NSS-I database: RCHJ.N
and 12, The variable RCHJJM is considered to be a "map" attribute, measured during the site selec-
tion process on 1:24,000-scale United States Geological Survey (USGS) topographic maps. This vari-
able is a measure of the length of reach between the intended sampling points (originally mapped
reach ends). In contrast, the variable L2 is a measurement of the distance between the exact
sampling points actually visited by field crews. In general, the L2 measurement is shorter than
RCH_LN, because field crews very seldom were able to visit a reach at the exact spot originally
identified. RCH_LN is the variable used to make NSS-I reach length population estimates. RCHJJXl is
a measure of the length of each reach in the NSS sample. The total reach length represented in the
NSS target population can be estimated by summing, for all reaches, the product of RCH_LN and the
sample weight, W.
NSS-I population estimates of stream length can be generated in several ways. One method
assigns the chemistry of one reach end (upstream or downstream) to the entire reach length. A
second method interpolates chemistry along a reach based on the chemistry of both reach ends.
Note: Interpolated length estimates can be made only by using the observations of reaches from
which water samples were collected at both upstream and downstream ends. It is important to pay
attention to drop code values when subsetting data for making population estimates of reach length.
Drop code (DRPCDE) < 1 identifies the set of observations for which both the upstream and down-
stream reach ends were sampled, in contrast to DRPCDE < 1, which identifies estimates for all target
observations, regardless of whether both reach ends were sampled.
48
-------
A.8 TOPOGRAPHIC DRAINAGE AREA MEASUREMENTS: av a2, %, a4, %, a,^, (A1, A2, A3, A4, A5,
AND A_WS)
(Detailed In Kaufmann et al., 1988; Section 2.4.1)
There are several different measurements of topographic drainage area in the NSS-I database.
Each reflects a different component of reach topographic drainage area The first, A1, is a measure-
ment of the direct drainage area contributing to the length of reach designated by the mapped
location of reach ends (upstream and downstream ends) identified during the site selection process.
This measurement is based on the location of reach ends, as identified on 1:250,000-scale maps
during the site selection process. These locations were then transferred to 1:24,000-scale maps on
which the drainage areas were measured.
The topographic drainage areas of NSS upstream ends are contained in the variables A2 and A3.
For nonheadwater reaches, A2 is a measure of the drainage area contributing flow to the upstream
reach end (measured on 1:24,000-scale maps). This value will be zero for headwater reaches (those
observations with RCH_HW=1). For headwater reaches, A3 is a measure of the drainage area
contributing flow to the upstream reach end. This value will be zero for nonheadwater reaches (i.e.,
those with RCHJHW > 1).
The variable A_WS is the total topographic drainage area contributing to stream flow at the
mapped sampling point (reach end) and is the drainage area variable used in making NSS-I
population estimates of watershed area. A_WS is the total drainage area contributing to a reach end.
As with A1, A_WS is based on the "map" location of reach ends identified during the site selection
process. Whereas the variable A1 is a measure of the direct drainage area to the intended
downstream point on a reach, the variable A_WS is equal to the sum of A1 and A2 for nonheadwater
reaches. A3 is incorporated into A1 for headwater reaches. Thus, for headwaters, A_WS = A1. For
downstream observations (where NODE=L), A_WS is equal to the drainage of the entire reach,
calculated as:
A_WS = A1 + A2: for nonheadwaters, or
A_WS = A1: for headwaters.
For upstream observations (where NODE = U), A_WS is calculated as :
A_WS = A2: for nonheadwaters, or
A_WS = A3: for headwaters.
Note: In contrast to A1, a measure of the direct drainage area of a mapped reach (as identified on
1:250,000-scale maps), A4 is a measure of the reach drainage area that drains the area between the
exact upstream and downstream field sampling locations. In turn, the variable A5 is a measure of the
49
-------
topographic drainage area that contributes to the exact upstream field sampling location (i.e., the
drainage area above the upstream sampling point).
Total drainage areas, based on the exact locations where water samples were collected
(analogous to A_WS), can be calculated as the sum of the variables A4 and A5, and are usually
slightly different than the A_WS measurements. This type of drainage area measurement may be
useful for data analyses that require a more precise association between water chemistry and
drainage area (e.g., examination of conductivity versus drainage area).
The units for the variables A1, A1 PRIME, A2, and A3 are mi2, whereas the units for the variables
A4, A5, and A_WS are km2. The conversion factor of 2.59 was used to convert mi2 to km2. NSS
population estimates of watershed area were calculated in km2.
A.9 SYNTHESIZED DATA FOR THE SOUTHERN BLUE RIDGE
(Detailed In Section 3.5.2)
Each observation in NSSIDS4 represents an actual water sample collected during the NSS-I or
Pilot Survey. In the Southern Blue Ridge Province (identified as subregion 2AS in NSSIDS4), only 20
of the 54 probability reaches were sampled at both upstream and downstream sampling locations
during the spring base flow period. In order to generate population estimates compatible with other
NSS-I subregions (in which equal numbers of upstream and downstream sites were sampled), a
supplemental data set was synthesized for 22 chemical parameters for the 34 upstream sites that
were not sampled. This information was generated from regression relationships among sampled
streams and from data collected during a summer sampling period. The rationale and equations used
to generate the data are presented in Kaufmann et al. (1988), Appendix B. This synthesized
information is essential for replicating reported NSS-I population estimates of upstream nodes in the
Southern Blue Ridge Province. To use these data, data set SBRSYN should be appended to
NSSIDS4. Caution should be used, however, in applying this information to trend analysis or
multivariate examination of data, since the data were generated from regression relationships of data.
A. 10 DIFFERENCES BETWEEN THE NSS-I AND THE NSS-I PILOT SURVEY
(Detailed In Kaufmann et al., 1988; Section 3.9)
Although most of the sampling and analysis methods used in the NSS-I main survey were
developed during the Pilot Survey, conducted a year earlier, some specific differences in the methods
used in the NSS-I should be noted, for example, phosphorus analysis, sample holding times, methods
of fractionation, and determination of aluminum species.
50
-------
A.11 REVISIONS TO NSS-I PILOT SURVEY DATA
(Detailed In Appendix C)
After the release of the NSS-I Pilot Survey report (Messer et al., 1988), revisions were made to a
small number of variables in the database. These revisions are discussed in detail in Appendix C.
The principal updates include the revision of one direct drainage measurements (A1) and revised ANC
and BNC measurements based on the refinement in the calculation algorithm. The one A1 change is
important because A1 is used in calculating the final sample weight for each stream. The estimated
total number of target reaches in the Southern Blue Ridge Province changed by 10 reaches.
Revisions to ANC and BNC were made on all observations in the NSS-I Pilot Survey database. The
original Pilot Survey data, as presented in the Pilot Survey report, are contained in PILOTDS3 and
PILOTDS4. The revised data are included in NSSIDS4 as data for subregion 2As. the Southern Blue
Ridge Province.
A. 12 COMPARISON OF PARAMETER UNITS IN THE NSS-I AND THE NLS
In the development of the databases associated with the National Surface Water Survey, there
was a conscious effort to maintain consistency in database content and variable format. There are,
however, some differences in the units maintained in the ELS and WLS databases and the NSS-I. The
following is a brief comparison of principal parameters measured in the NSWS. Units of calculated
parameters are maintained in the units of their component variables (e.g., the unit for sum of anions is
Parameter(s)
Anions/Cations
ANC/BNC
Conductivity
A|2+ CQ kiln D
Ml , re, Mil, r
Si02
Watershed Area
DOC, DIC
a inrliratfis narar
NSS-I
Units Units
Measured Converted1
mg L"1 |ieq L"1
neq L'1
(iS cm"1
* -j
km2 or mi2
mg L-1
neters maintflinpH in nrininallu moaci
NLS
Units Units
Measured Converted1
mg L"1 ^eq L"1
neq L'1
liS cm"1
i "i
riJ >-
mg L'1
ha
mg L'1
irorl units:
51
-------
In the NSS-I, parameters calculated from pH and DIG (e.g., HCO3" CO32", H+, OH", A") are based
on the variables PHSTVL and DICVAL (measured at the analytical laboratory on closed-headspace
samples). In contrast, in the NLS, these parameters are calculated from the variables PHAC11 and
DICI11 (measured at the analytical laboratory). Section A.16 details the reasoning for using these
specific pH and DIG measurements.
A. 13 REVISIONS TO a1
(Detailed in Appendix D)
After the release of the NSS-I final report (Kaufmann et al., 1988), several direct drainage
measurements (a^ were revised. Any changes in a1 are important because they are used in
generating each reach's final sample weight If a substantial weight change occurs, a subsequent
change will occur in the estimated total number of sites, which in turn may alter the reported
estimated number of reaches within a specific chemical range. Examination of the impact of revising
the a1 and sample weights indicates that a significant change in population distributions does not
occur and therefore the conclusions presented in the NSS-I final report (see Appendix E) do not
change. The distributed NSS database includes the data necessary to replicate any estimates
presented in the NSS-I report, as well as the revised a1 measurements. The variable A1 contains the
a1 measurement used to generate population estimates, whereas the variable A1 PRIME contains the
revised a1 values. Note: if a revision to a1 did not occur, then A1 = A1 PRIME.
A. 14 NSS DATABASE VARIABLE FORMATS
All NSS database values were recorded in the format in which they were reported by the
analytical laboratory, processing laboratory, or field sampling crew. Table B-1 in Appendix B shows
the SAS variable formats used in printouts of data listings. For both calculated and reported chemical
values, printed decimal places do not necessarily represent actual precision. Results of NSS-I
chemistry precision assessments are reported in Cougan et al. (1988) and in Section 4 of Kaufmann
et al. (1988). In general, these reporting formats maintain a decimal place beyond the intended level
of interest for the data. No data values were rounded off before or during data entry. Variations in
reporting precision are most likely to result from the format in which values were originally recorded by
the analytical laboratory, processing laboratory, or field sampling crews. Values are stored in SAS
data sets using floating point precision. Calculated variables, including those with a '16' suffix (e.g.,
SO416), are not rounded off in the database, but maintain the same reporting format as the variables
that were used to calculate them (e.g., SO411).
52
-------
A.15 SUBREGION IDENTIFICATION CODES
The variables RCHJD and STRMJD uniquely identify each sampling location. Both variables
contain concatenated information for reach subregion and a 1:250,000-scale map identification code
(from maps used in the site selection process). Note: The first two characters of these variables,
which indicate general mapped subregion areas, may not accurately identify the appropriate
subregion for a given reach.
The variable SUBJD refers to specific subregions (e.g., 1D, 2BN, 2X, etc.) for which population
estimates are made. For example, reaches that have SUBJD = "2X" include those that may have the
first two characters of RCHJD or STRMJD equal to "2A", "2B", or "2C". This group of reaches is from
a combined area of the Southern Appalachians, the southern area of the Valley and Ridge Province,
and a northern portion of the Blue Ridge Mountains. Refer to Figure 2-1 for the specific geographic
areas within each subregion. It is also important to note that the NSS is not based on the same
region and subregion numbering scheme as the NLS. For example, subregion "2D" in the NSS is in
the Ozarks and Ouachitas of Arkansas and Oklahoma, while subregion "2D" in the NLS is in the Upper
Peninsula of Michigan, Wisconsin, and Minnesota.
Streams in the following subregions may have different values for the SUBJD and STRMJD
variable prefix:
SUBJD STRMJD/RCHJD
Value Prefix Geographic Area
2X 2A Northern portion of the Blue Ridge
2X 2B Southern Appalachians
2X 2C Southern portion of Valley and Ridge
2As 2A NSS Pilot Survey area sampled in 1985. Observations for
these sites also have a shorter STRMJD and RCHJD than do
those from the full-scale survey
2BN 2B Northern portion of the Valley and Ridge Province
2CN 2C Northern Appalachians
A.16 USING DIG AND pH IN CALCULATED VARIABLES
Several calculated variables in the NSS database(s) incorporate measurements of dissolved
inorganic carbon (DIG) and pH (e.g., HCO316, CO316, H16, OH16, ANSUM, and CATSUM). In the
NSS, as with other NSWS surveys, these parameters were originally calculated from measurements
made at the analytical laboratory (i.e., database variables DICI11 and PHAC11). These specific
53
-------
variables were used because they were considered to represent the chemistry of the cubitainer water
sample (from which other parameters are measured such as anions and cations). Because NSS
water samples were often over-saturated with respect to C02, it is important that both pH and DIG
were measured at the same time so that any changes in DIG are reflected in measured pH.
After the release of the NSS-I report, re-examination of calculated variables (e.g., HCO316,
C0316, etc.) indicated that DIG (DICI11) may not always have been measured close enough in time to
pH (PHAC11) to avoid CO2 concentration changes from degassing between the two measurements.
Examination of verification flags identified approximately 25% of NSS-I routine water samples as not
having initial DIG measurements made within the recommended 14-day holding time, while only about
1% of the corresponding pH measurements were not measured within the recommended 14-day
holding time. Based on this information, it is recommended that calculated variables which
incorporate DIG or pH measurements use those made at the processing laboratory (i.e., database
variables DICVAL and PHSTVL). These (closed-system) syringe sample measurements minimize any
CO2 degassing prior to pH or DIG measurement. In addition, these samples were "processed" within
24-36 hours after the sample was collected from a stream. Because DICVAL and PHSTVL are
considered to provide a "matched" set of DIG and pH measurements, they were used to calculate
components of the carbonate buffering system (HCO316, CO316, OH16, and H16). Please note,
however, that these measurements were not made on the same cubitainer subsample as the anions
and cations.
The variables HC0316, CO316, OH16, and H16 in the NSS-I database have been revised
accordingly (using DICVAL and PHSTVL) and will therefore be slightly different than those presented
in the NSS-I report volumes.
54
-------
APPENDIX B
NSS-I DATA DICTIONARY
The following data dictionary describes the contents of the U.S. Environmental Protection
Agency's (EPA) National Stream Survey (NSS) Pilot and Phase-l database. This dictionary is provided
to aid data managers, programmers, and users of the NSS database in the accurate transfer and use
of the NSS data on their own computer systems. EPA methods referenced in this appendix are taken
from U.S. EPA (1983); USGS methods are from Skougstad et al. (1979). When appropriate, the
method or equation used to generate each variable is discussed or referenced. Additional method
descriptions are discussed in Sections 2 and 3 of Kaufmann et al. (1988).
Detailed protocol descriptions are presented in the following documents:
Field Sampling Protocols
Pilot Survey Knapp et al. (1987)
NSS-I Hagleyetal. (1988)
Processing Laboratory Protocols
Pilot Survey - Knapp et al. (1987)
NSS-I - Arentet al. (1988)
Analytical Laboratory Protocols
Pilot Survey - Drouse et al. (1986, 1987)
NSS-I - Hillman et al. (1987); Cougan et al. (1988)
55
-------
TABLE B-1. NSS-I DATA DICTIONARY
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
ACCO11 iieqL'1 Num 9.3
ALDSVL mg L'1 Num 9.3
ALDS16 jimol L'1 Num 9.3
ALEX11 mg L"1 Num 9.3
ALEX16 jimolf1 Num 9.3
ALINOR (imol L'1 Num 8.3
ALKA11 jieq L'1 Num 9.3
Acidity, or base neutralizing capacity, measured at the
analytical laboratory; a measure of the amount of base
needed to neutralize carbonate species, hydronium, and
other acids in the sample. Determined by Gran analysis of
base titration data (Hillman et al., 1987; Kramer, 1984).
An estimate of total monomeric aluminum measured by
pyrocatechol violet (PCV) colorimetry and automated flow
injection analyzer at the processing laboratory, not measured
in the Pilot Survey (Hillman et al., 1987).
ALDS16 = ALDSVL * 1000 / 26.982
Extractable aluminum, an estimate of monomeric aluminum
complexes (AI3+). A filtered unacidified sample was
complexed with 8-hydroxyquiniline, extracted with methyl-
isobutyl ketone (MIBK), and analyzed by Graphite furnace
(GF) atomic absorption spectroscopy (AAS) at the analytical
laboratory (Hillman et al., 1987; EPA method 202.2).
ALEX16 = ALEX11 * 1000 / 26.982
Calculated inorganic monomeric aluminum.
ALINOR = ALEX16 - ALOR16: in NSS-I Pilot Survey (based
on MIBK aluminum methods).
ALINOR = ALDS16 - ALOR16: in NSS-I (based on PCV
methods).
Note: Negative values have been scored to 0.
Alkalinity, or acid neutralizing capacity (ANC); a measure of
the amount of acid necessary to neutralize the bicarbonate,
carbonate, hydroxyl, and other bases in the sample;
measured at the analytical laboratory by Gran analysis of acid
titration data (Hillman et al., 1987; Kramer, 1984).
56
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
ALORVL mg L'1 Num 9.3
ALOR11 mg L"'
ALTL16
Num 9.3
ALOR16 (imol L'1 Num 9.3
ALTL1 1 mg L'1 Num 9.3
ANDEF jieq L"1 Num 9.3
An estimate of nonextractable monomeric (organic) aluminum
measured at the processing laboratory using the MIBK
method after passing the sample through a strong cation
exchange column. Measured only in Phase I samples (not in
Pilot Survey) Hillman et al., 1987).
Nonextractable organic monomeric aluminum measured at
the processing laboratory using the PVC method after
passing the sample through a strong cation exchange
column. Measured only in Pilot Survey (EPA method 202.2).
ALOR16 = ALORVL * 1000/26.982: in Phase I.
ALOR16 = ALOR11 * 1000/26.982: in Pilot Survey.
Total aluminum measured on an unfiltered, acidified (HNOg)
aliquot at the analytical laboratory after digestion; analyzed by
graphite furnace AAS (Hillman et al., 1987; EPA method
202.2).
L'1 Num 9.3 ALTL16 = ALTL11 * 1000/ 26.982.
Anion deficit, or total cations (CATSUM) minus total anions
(ANSUM).
ANSUM (ieq L"1 Num 9.3 Total anions, defined as:
ANSUM = HCO316 + CO316 + CL16 + NO316 + S0416
+ FTL16 + OH16
A1
mi2 Num 7.3 Direct drainage area. The portion of the watershed that
drains directly into a reach between the upstream and
downstream confluences interpreted from topographic maps.
The variable A1 is used in calculating sample weights that are
used in making statistical estimates of the NSS-I target
population. Measurements of A1 are maintained in their
measured units and not converted to metric values for
reasons of statistical protocol. Measured on 1:24,000-scale
USGS topographic maps.
57
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
A4
krrf
A5
knf
A WS km2
BAT ID
Num 7.3 The watershed area draining directly to the segment of
stream between the upstream and downstream points where
water samples were collected.
Note: while conceptually similar, the variables A1 and A4 will
not necessarily be equal as field crews were seldom able to
collect water samples at the exact locations of the upstream
and downstream ends of the mapped reach. Measured on
1:24,000-scaie USGS topographic maps.
Num 7.3 The watershed area draining to the upstream site where water
samples were collected.
Num 7.3 The total watershed area contributing to stream flow at a
mapped reach end (upstream or downstream). This is the
drainage area measurement used in making NSS population
estimates. This measurement is based on the location of
reach ends, as identified on 1:250,000-scale maps prior to
sampling. These locations were then transferred and
drainage areas were measured on 1:24,000-scale USGS
topographic maps.
Note: The variable A_WS is not sum of A4 and A5. Topo-
graphic drainage area based on the actual locations where
water samples were collected, calculated using the variables
A4 and A5, are usually slightly different than the A_WS
measurements (see Appendix A, Section A. 8).
Char 6 Identification code for a batch (group) of samples assigned at
the processing laboratory. The combination of BATJD and
SAMJD form a unique sample identifier.
CATSUM jieq L"1 Num 9.3 Total cations, defined as:
CATSUM = CA16 + MG16 + K16 + NA16 + NH416 + H16.
58
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
CA1 1
CA16
mg L"
-1
Num 9.3 Calcium measured at the analytical laboratory on an acidified,
filtered aliquot using flame AAS or ICPES (EPA method
215.1).
Num 9.3
CA16 = CA11 * 49.90.
CL11
mg L"
CL16
Num 9.3 Chloride ion measured at the analytical laboratory on a
filtered unacidified aliquot using ion chromatography (ASTM,
1984; O'Dell et al., 1984).
Num 9.3 CL16 = CL11 * 28.21.
COLVAL Platinum Num 8.0
Cobalt
Units
COND11 jiS cm"1 Nurn 9.3
CONIS
COUNTY1
COUNTY2
COUNTY3
cm"1 Num 9.3
CONVAL uS cm"1 Num 9.3
True color measured after centrifuging at the processing
laboratory using a Hach Model CO-1 Color Test Kit (EPA
method 110.2 modified).
Specific conductance, temperature corrected, measured at
the analytical laboratory (EPA method 120.1).
Specific conductance measured in situ by field crews.
Temperature corrected in Phase I, but not corrected in the
Pilot Survey (Hagley et al., 1988).
Specific conductance, temperature corrected, measured at
the processing laboratory. Measured only in Phase I survey.
Char 20 Name of the county where the sample site and watershed are
located. Identified on 1 :24,000-scale USGS topographic
maps.
Char 20 Name of second county, if necessary, where the sample
watershed is located. Identified on 1 :24,000-scale USGS
topographic maps.
Char 20 Name of third county, if necessary, where the sample
watershed is located. Identified on 1:24,000-scale USGS
topographic maps.
59
-------
TABLE B-1. NSS-! DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
COUNTY4
C0316
DATSMP
DICE11 mg L'1
DICVAL mg
DOJS mg [_''
Char 20 Name of fourth county, if necessary, where the sample water-
shed is located. Identified on 1:24,000-scale USGS
topographic maps.
Num 9.3 Calculated carbonate concentration
CO316 = 60009 * (DICVAL/12011) * ALPHA2 * 33.33
where:
ALPHA2 = K1*K2 / ((H**2) + (H * K1) + (K1*K2))
where:
K1 = 4.463 x 10**-7
K2 = 4.6881 x 10**-11
H = 10**(-PHSTVL)
(See Appendix A, Section A. 16)
Num DATE? Date of sampling visit by field crew.
Num 9.3 Air-equilibrated dissolved inorganic carbon measured at the
analytical laboratory using a carbon analyzer on an unfiltered,
unacidified aliquot that was purged for 20 minutes with 300
ppm C02 air (also, see PHEQ11) (EPA method 415.2,
modified).
DICI11 mg L'1 Num 9.3
Dissolved inorganic carbon measured at the analytical
laboratory using a carbon analyzer on an unfiltered,
unacidified aliquot (EPA method 415.2, modified).
Num 9.3 Dissolved inorganic carbon measured at the processing
laboratory using a carbon analyzer on a closed-headspace
sample collected and transported in a 60-ml syringe (EPA
method 415.2, modified).
Num 8.1 Dissolved oxygen measured in situ at the field sampling site
(Hagley et al., 1988).
60
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
DOC11 mg L"1 Num 9.3
Dissolved organic carbon measured at the analytical
laboratory in a filtered, acidified (H2SO4) sample using a
carbon analyzer (EPA method 415.2).
DRPCDE
Num 2.0 Numeric code used to indicate whether a specific observation
is considered interest or noninterest in the target population
of reaches represented in the NSS. This variable was created
to allow the selective exclusion of noninterest observations in
statistical analyses. These values are encoded as follows:
Drop
Code
13
Exclusion Criteria Description
NSS-I target observation for both the upstream and
downstream sample site.
Alternate node of noninterest site (useful in analyses
which require a data set of reaches for which both
the upstream and downstream sites were sampled,
e.g., estimates of interpolated reach length).
Noninterest sites (Kaufmann et al., 1988; Section
3.2). Observation not included in NSS-I target
population estimates (because of tidal influence,
intermittent flow conditions, high conductivity, etc.)
Site is acidic due to acid mine drainage (Kaufmann
et al., 1988; subsection 9.3.1).
Pilot Survey nonspring index data (water samples
not collected between March 15 and May 15)
(Kaufmann et al., 1988; subsection 1.3.2.1).
NSS-I special interest sites (Kaufmann et al., 1988;
Section 2.6).
Combination of DRPCDE 1 and 3 (i.e., an alternate
node of a noninterest site and is impacted by acid
mine drainage).
Note: Values with DRPCDE values > 1 were excluded when
generating estimates of the NSS-I target population presented
in the final NSS-I report (Kaufmann et al., 1988).
61
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS Suggested
Variable Units/ Format
Name Format Type Width Variable Definition
ELEV m
Num 7.2
FE11 mg L'1 Num 9.3
FE16
FTL1 1 mg L'1 Num 9.3
FTL16
GRADE %
L'1 Num 9.3
Num 8.2
HC0316 (ieq L'1 Num 8.3
Elevation of visited sampling site. Measured on 1:24,000-
scale USGS topographic maps.
Dissolved iron measured at the analytical laboratory on a
filtered, acidified aliquot using flame AAS or inductively
coupled plasma emission spectroscopy (ICPES) (EPA method
236.1).
L'1 Num 9.3 FE16 = FE11 * 1000 / 55.84
Total dissolved fluoride measured at the analytical laboratory
in a filtered, unacidified aliquot using an ion-sensitive
electrode (ISE) (EPA method 340.2, modified).
FTL16 = FTL11 * 52.64
Reach gradient; based on the difference in elevation and
distance between the upstream and downstream sampling
sites. From measurements made on 1:24,000-scale USGS
topographic maps.
upstream sample downstream sample
GRADE = site elevation (m) site elevation (m)
sampled length (m)
Sampled length = L2 * 1000
Calculated bicarbonate concentration:
HC0316 = 61017 * (DICVAL/12011) * ALPHA1 * 16.39
where:
ALPHA1 = (H * K1) / ((H**2) + (H * K1) + K1*K2))
where:
K1 = 4.4463 x 10**-7
K2 = 4.6881 x 10**-11
H = 10**-PHSTVL
(See Appendix A, Section A. 16)
62
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
H16 !ieq L
-1
K11
mg L"'
K16
LABNAM
LAT_STD Decimal Num 9.4
Degrees
LON_STD Decimal Num 9.4
Degrees
L2
km
MAPI
MAP2
MAPS
MAP4
MAPS
Num 9.4 Calculated hydronium concentration.
H16 = (10**-PHSTVL) * (10**6)
Num 9.3 Dissolved potassium measured on a filtered, acidified aliquot
at the analytical laboratory using flame AAS (EPA method
258.1).
Num 9.3 K16 = K11 * 25.57
Char 20 Name of the analytical laboratory that performed the analyses.
GLOBAL = Global Laboratories
NYSDOH = New York State Department of Health
Latitude of the visited sample site. Measured from 1:24,000-
scale USGS topographic maps.
Longitude of the visited sample site. Measured from
1:24,000-scale USGS topographic maps.
Num 9.3 Length of the reach segment between the visited upstream
and downstream field sampling sites (see also RCHJJxl).
Measured from 1:24,000-scale USGS topographic maps. L2
was not used in making NSS population estimates of length.
Char 32 Name of the 1:24,000-scale USGS map showing the sample
site and watershed location.
Char 32 Name of the second 1:24,000-scale USGS map, if necessary,
showing the sample watershed location.
Char 32 Name of the third 1:24,000-scale USGS map, if necessary,
showing the sample watershed location.
Char 32 Name of the fourth 1:24,000-scale USGS map, if necessary,
showing the sample watershed location.
Char 32 Name of the fifth 1:24,000-scale USGS map, if necessary,
showing the sample watershed location.
63
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS Suggested
Variable Units/ Format
Name Format Type Width Variable Definition
MAP6
MN16
NODE
Char 32
MG1 1 mg L'1 Num 9.3
MG16 (ieq L'1 Num 9.3
MN11 mg L'1 Num 9.3
Name of the sixth 1 :24,000-scale USGS map, if necessary,
showing the sample watershed location.
Dissolved magnesium determined on a filtered, acidified
aliquot at the analytical laboratory using flame AAS or ICPES
(EPA method 242.1).
MG16 = MG11 * 82.26
Dissolved manganese measured at the analytical laboratory
using flame AAS or ICPES on a filtered, acidified aliquot (EPA
method 243.1).
L1 Num 9.3 MN16 = MN11 * 1000 / 54.938.
NA1 1 mg L/1 Num 9.3
Dissolved sodium measured on a filtered acidified aliquot at
the analytical laboratory using flame AAS (EPA method
273.1).
NA16 jieq L'1 Num 9.3 NA16 = NA11 * 43.50.
NH411 mg L'1 Num 9.3
Ammonium measured at the analytical laboratory in a filtered,
acidified (HgSO^ aliquot using automated colorimetry
(phenate) (EPA method 350.1).
NH416 jieq L'1 Num 9.3 NH416 = NH411 * 55.44.
Char 1.0 The variable used to identify sampling locations or sample
sites on a given reach. These locations are identified as
either upstream (U) or downstream (L) ends or nodes. The
point below the confluence of two reaches defines an
upstream node and the point above the confluence of two
reaches defines the downstream node.
Note: Water samples were collected from an area of the
reach considered well mixed when sampling below a con-
fluence.
64
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS Suggested
Variable Units/ Format
Name Format Type Width Variable Definition
NOTSAM
Char 30 Reason, when appropriate, why a stream was not sampled.
Based on comments and decisions of field sampling crews.
NO31 1 mg L'1 Num 9.3
NO316 jieq L'1 Num 9.3
OH16 neq L'1 Num 9.3
ORGION neqL'1 Num 9.3
PH_CLO pH units Num 4.2
PH_R pH units Num 4.2
PHAC11 pH units Num 4.2
Nitrate ion measured at the analytical laboratory on a filtered,
unacidified aliquot using ion chromatography (ASTM, 1984;
O'Delletal., 1984).
NO316 = NO311 * 16.13.
Calculated Hydroxide ion concentration:
OH16 = (10**(PHSTVL-14)) * (10**6).
Estimated organic anion concentration (Oliver Model; Oliver,
1983).
ORGION = [(10 **(-PK)) * DOC11 * 10] / [(10**(-PK) +
(10**-PHAC11)]
where
PK = 0.96 + 0.9 * PHAC11 0.039 * (PHAC11 ** 2)
Measured only in NSS-I Pilot Survey. Sample pH measured
at the field sampling site in a closed container with a portable
pH meter (Beckman pHI-21) and glass combination electrode
(Orion Ross Model 8104) (Knapp et al., 1988). Measured only
in Pilot Survey.
Sample pH measured at the field sampling site in an open
container with a portable meter (Beckman pHI-21) and glass
combination electrode (Orion Ross Model 8104) (Hagley et
al., 1988).
Initial pH from the acidity titration prior to the addition of base
titrant. Measured at the analytical laboratory on an unfiltered,
unacidified aliquot stirred in a CO2-free vessel (EPA method
150.1).
65
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS Suggested
Variable Units/ Format
Name Format Type Width Variable Definition
PHAL11 pH units Num 4.2
PHEQ11 pH units Num 4.2
PHSTVL pH units Num 4.2
PTD11
mg L"
-1
PTL16
QUAD
RCH HW
Num 9.3
PTD16
PTL11
mgL'1
Num 9.3
Num 9.3
Num 9.3
Char 30
Num 2.0
Initial pH from the alkalinity titration prior to the addition of
acid titrant. Measured at the analytical laboratory on an
unfiltered, unacidified aliquot, in an open (exposed to air)
vessel (EPA method 150.1).
Air-equilibrated sample pH measured at the analytical labora-
tory on an unfiltered, unacidified aliquot aerated with 300 ppm
CO2 air for 20 minutes (EPA method 150.1).
Sample pH measured at the processing laboratory with a por-
table pH meter (Beckman pHI-21) and glass combination
electrode (Orion Ross Model 8104) on a closed-headspace
sample. Variable used to make NSS population estimates of
pH (EPA method 150.1).
Total dissolved phosphorus measured at the analytical labora-
tory using automated colorimetric (phosphomolybdate)
methods. Determined on a filtered aliquot (measured only in
Phase I survey) (USGS 1-4600-78, modified).
PTD16 = PTD11 * 1000/30.974
Total phosphorus measured at the analytical laboratory using
automated colorimetric (phosphomolybdate methods). Deter-
mined on an unfiltered aliquot (measured only in the Pilot
Survey) (USGS 1-4600-78, modified).
PTL16 = PTL11 * 1000 / 30.974
Name of the 1:250,000-scale USGS topographic map on
which the sample site appears.
Number of headwater reaches upstream of the mapped
sample location on 1:250,000-scale maps (Shreve order from
1:250,000-scale maps).
66
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
RCH ID
RCH LN km
SAM ID
SAMCOD
Char 8 Reach identification code (e.g., 3B041016). An 8-digit code
containing three fields that indicate the (1) NSS-I subregion
(3B), the 1:250,000-scale map ID (041), special interest site
designation (0 = Routine sites, 9 = Special interest sites),
and the grid-dot number (16).
Note: The first two characters of this variable, which indicate
mapped subregion areas, may not accurately identify the
appropriate NSS-I subregion for a given reach (see Appendix
A, Section A. 15).
Num 8.3 Length of the mapped reach segment between the upstream
and downstream ends (associated with site selection
process) identified on 1:250,000-scale maps, but measured
on 1:24,000-scale USGS topographic maps. Variable used in
making NSS population estimates of reach length.
Note: This value is the mapped length of the reach as origin-
ally mapped and not a measure of the distance between vis-
ited sampling points. It is not unusual for this value to be
different from L2, the distance between visited sampling
points.
Char 3.0 Sample identification code within a batch, assigned at the
processing laboratory. The combination of BATJD and
SAMJD form a unique sample identifier.
Char 3.0 Sample type code, or combination of codes, used to identify
the sample as follows (note that precipitation event codes, E,
are always listed first, if present):
D = Duplicate sample
DA = Routine/Duplicate average
E = Suspected precipitation event influence (episode)
EDA = Routine/Duplicate average influenced by precipitation
event
ER = Routine sample influenced by precipitation event
R = Routine
SY = Synthesized chemistry values
67
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
SAMRN
SHRE75
Num 1.0 Sample visit number identifying the visit number to the
sample site.
Num 3.0 Shreve stream order measured from 1:24,000-scale maps.
SIO211 mg L'1 Num 9.3
Silica measured at the analytical laboratory using automated
colorimetry (molybdate blue) (USGS 1-2700-78).
SI0216 jimol L'1 Num 9.3
SIO216 = SIO211 * 1000 / 60.084
SIT CLS
Char 6.0 Reach or observation noninterest code. Used to distinguish
impacts that result in a sampling location being considered
noninterest or nontarget.
SOBC
jieq
Num 9.3
A1 = Acid mine drainage affects upper node.
A2 = Acid mine drainage affects lower node.
A3 = Acid mine drainage affects both nodes.
C1 = Conductance greater than 500 jiS/cm at upper node
C2 = Conductance greater than 500 jiS/cm at lower node
C3 = Conductance greater than 500 (iS/cm at both nodes.
11 = Intermittent flow at upper node.
12 = Intermittent flow at lower node.
13 = Intermittent flow at both nodes.
03 = No evident channel at both nodes (swamp, or lake).
R = Random sample miss; no alternate node, reach dropped.
S1 = Special case at upper node.
S2 = Special case at lower node.
S3 = Special case at both nodes.
T3 = Both nodes tidally influenced.
Note: A given record may have multiple values. When
multiple noninterest conditions occurred on a reach, these
codes are combined, (i.e., high conductivities found at both
sample sites in addition to acid mine drainage evidence at a
downstream end would be classified as "A2C3").
Sum of base cations.
SOBC = NA16 + K16 + MG16 + CA16
68
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS
Variable Units/
Name Format
Suggested
Format
Type Width Variable Definition
SO411 mg L'1
SO416 jieq L"1
STATE1
STATE2
STRATUM
STRA75
STRA250
STRM ID
STRMDP m
STRMNAM
Num 9.3 Sulfate measured at the analytical laboratory using ion
chromatography (ASTM, 1984; O'Dell et al., 1984).
Num 9.3 SO416 = SO411 * 20.82
Char 2 Two-letter postal abbreviation for the state in which the
sample site and watershed are found. Identified on 1:24,000-
scale USGS topographic maps.
Char 2 Two-letter postal abbreviation for a second state, if any, in
which the sample watershed is found. Identified on 1:24,000-
scale USGS topographic maps.
Num 1.0 Statistical stratum (from the site selection process) for the
reach (Section 2.3.2).
1 = Regular
2 = Low ANC
3 = Small a1
Num 3.0 Strahler order measured from 1:24,000-scale maps (Strahler,
1957).
Num 2.0 Strahler order measured from 1:250,000-scale maps (Strahler,
1957). Measured only in Phase-l Survey.
Char 9.0 Site identification code concatenates RCHJD and NODE.
Note: The first two characters of this variable, which indicate
mapped subregion areas, may not accurately identify the
appropriate NSS-I subregion for a given reach (see Appendix
A, Section A. 15).
Num 3.1 Representative depth at field sampling site.
Char 30 Stream name.
69
-------
TABLE B-1. NSS-I DATA DICTIONARY (continued)
SAS Suggested
Variable Units/ Format
Name Format Type Width Variable Definition
STRMWD m Num 4.1 Representative stream width at field sampling site.
SUBJD Char 3.0 Subregion identification code (See Appendix A, Section A. 15):
1D = Poconos/Catskills
2CN = Northern Appalachians
2BN = Valley and Ridge
3B = Mid-Atlantic Coastal Plain
2AS = Southern Blue Ridge
3A = Piedmont
2X = Southern Appalachians
2D = Ozarks/Ouachitas
3C = Florida
SI = Special Interest Sites
TIMSMP HH:MM Num TIMES Time at which water samples were collected.
TMPSTR °C
TURVAL NTU
W
Num 9.3 Stream temperature measured in situ at time of sampling.
Num 9.3 Turbidity measured at the processing laboratory using a
nephelometer (EPA method 180.1).
Num 12.6 The statistical weighting factor used in making population
estimates calculated as:
W = WC * (64/Max(AI,0.2)).
WC
Num 12.6 Conditional stage II sampling weight.
70
-------
Table B-2. NSS-I DATABASE FLAGS
FLAGS GENERATED VIA ANION/CATION BALANCE EVALUATION:
AO Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to unknown cause.
A1 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to unmeasured
anions/cations not considered in %IBD calculation.
A2 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to anion (flag suspect
anion) contamination.
A3 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to cation contamination.
A4 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to unmeasured organic
protolytes (fits Oliver Model).
A5 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to possible analytical
error - anion concentration too high (flag suspect anion).
A6 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to possible analytical
error - cation concentration too low (flag suspect cation).
A7 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to possible analytical
error - anion concentration too low (flag suspect anion).
A8 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to possible analytical
error - cation concentration too high (flag suspect cation).
A9 Anion/Cation % Ion Balance Difference (%IBD) was outside criteria due to possible analytical
error - alkalinity (ANC) measurement.
FLAGS GENERATED VIA QC BLANK SAMPLE EVALUATION:
BO External (field) blank is above expected criteria for pH, DIG, DOC, specific conductance, ANC,
and BNC determinations.
B1 Internal (lab) blank is >2 x CRDL for DIG, DOC, and specific conductance determinations.
B2 External (field) blank is above expected criteria and contributed >20% to sample concentrations.
(This flag is not used for pH, DIG, DOC, specific conductance, ANC, and BNC determinations.)
71
-------
TABLE B-2. NSS-I DATABASE FLAGS (continued)
B3 Internal (lab) blank is >2 x CRDL and contributes >10% to the sample concentrations. (This
flag is not used for DIG, DOC, and specific conductance determinations.)
B4 Potential negative sample bias based on internal (laboratory) blank data.
B5 Potential negative sample bias based on external (field) blank data.
FLAGS GENERATED VIA CONDUCTANCE BALANCE EVALUATION:
CO % Conductance Difference (%CD) was outside criteria due to unknown cause.
C1 % Conductance Difference (%CD) was outside criteria due to possible analytical error - anion
concentration too high (flag suspect anion).
C2 % Conductance Difference (%CD) was outside criteria due to anion contamination.
C3 % Conductance Difference (%CD) was outside criteria due to cation contamination.
C4 % Conductance Difference (%CD) was outside criteria due to unmeasured organic ions (fits
Oliver Model).
C5 % Conductance Difference (%CD) was outside criteria due to possible analytical error in specific
conductance measurement.
C6 % Conductance Difference (%CD) was outside criteria due to possible analytical error - anion
concentration too low (flag suspect anion).
C7 % Conductance Difference (%CD) was outside criteria due to unmeasured anions/cations (other
anions/cations not considered in %CD calculation).
C8 % Conductance Difference (%CD) was outside criteria due to possible analytical error - cation
concentration too low (flag suspect cation).
FLAGS GENERATED VIA DUPLICATE PRECISION EVALUATION:
C9 % Conductance Difference (%CD) was outside criteria due to possible analytical error - cation
concentration too high (flag suspect cation).
D2 External (field) duplicate precision exceeded the maximum expected % Relative Standard
Deviation (%RSD), and both the routine and duplicate sample concentrations were greater than
10 times the Contract Required Detection Limit (CRDL).
72
-------
TABLE B-2. NSS-I DATABASE FLAGS (continued)
D3 Internal (lab) duplicate precision exceeded the maximum contract required % Relative Standard
Deviation (%RSD), and both the routine and duplicate sample concentrations were >10 x
Contract Required Detection Limit (CRDL).
FLAGS THAT IDENTIFY SUSPECT FIELD DATA:
FO % Conductance difference (%CD) exceeded criteria when in situ field conductance value was
replaced.
F1 Hillman/Kramer protolyte analysis program indicated field pH problem when stream site pH value
was replaced.
F2 Hillman/Kramer protolyte analysis program indicated unexplained problem with stream site pH or
processing laboratory DIC values when stream site pH value was replaced.
F3 Hillman/Kramer protolyte analysis program indicated field problem - processing laboratory pH.
F4 Hillman/Kramer protolyte analysis program indicated field problem - processing laboratory DIC.
F5 Hillman/Kramer protolyte analysis program indicated unexplained problem with processing
laboratory pH or DIC values when processing laboratory pH value was replaced.
F6 % Conductance Difference (%CD) exceeded criteria when processing laboratory (trailer) specific
conductance value was replaced.
FLAGS GENERATED VIA HOLDING TIME EVALUATION:
HO The maximum holding time criteria were not met.
H1 No 'Date Analyzed" data were submitted for reanalysis data.
FLAG GENERATED VIA DETECTION LIMIT EVALUATION:
L1 Instrumental Detection Limit (IDL) exceeded Contract Required Detection Limit (CRDL) and
sample concentration was <10 x IDL.
FLAG GENERATED VIA CONTRACT SPECIFICATION ASSESSMENTS:
MO Value obtained using a method that was outside criteria as specified by contract.
73
-------
TABLE B-2. NSS-I DATABASE FLAGS (continued)
FLAGS GENERATED VIA QC AUDIT SAMPLE EVALUATION:
NO Audit sample value exceeded upper control limit.
N1 Audit sample value was below the control limit.
FLAGS GENERATED VIA PROTOLYTE EVALUATION:
PO Lab problem - initial pH from alkalinity (ANC) titration.
P1 Lab problem - initial pH from acidity () determination.
P7 Lab problem - C02-Acidity (BNC) determination.
FLAGS GENERATED VIA QC CALIBRATION REFERENCE SOLUTIONS:
Q1 Quality Control Check Sample (QCCS) was above contractual criteria.
Q2 Quality Control Check Sample (QCCS) was below contractual criteria.
Q3 Insufficient number of QCCS were measured.
Q4 No Quality Control Check Sample (QCCS) was analyzed.
Q5 Detection Limit QCCS was not 2 to 3 times Contract Required Detection Limit (CRDL) and
measured value was not within 20% of the theoretical concentration.
74
-------
TABLE B-2. NSS-I DATABASE FLAGS (continued)
FLAGS GENERATED VIA DATA VALIDATION:
U1 Value considered error. Substitution made with observation from same stream.
U2 Value considered error. Substitution made based on known relationships with other variables.
W1 Unusual value in context of subregion chemistry, but reconcilable based on site chemistry.
W2 Unusual value in context of subregion and on-site chemistry, but not replaced.
FLAGS USED TO IDENTIFY VALUES FOUND DURING VERIFICATION:
XO Irreconcilable but confirmed based on QA review.
X1 Extractable Al concentration is greater than total Al concentration by 0.010 mg/L where
extractable Al > 0.015 mg/L
X2 Irreconcilable but confirmed data - possible aliquot switch.
X3 Irreconcilable but confirmed data - possible gross contamination of aliquot or parameter.
X4 Irreconcilable but confirmed data - possible sample (all aliquots) switch.
Values for flags XO through X4 should not be included in any statistical analysis.
75
-------
APPENDIX C
PILOT SURVEY REVISIONS
Results of the NSS-I Pilot Survey, conducted in the Southern Blue Ridge (NSS subregion 2As),
are presented by Messer et al. (1986,1988). The data are contained in PILOTDDS3 and PILOTDS4.
These results are also presented by Kaufmann et al. (1988), but with some minor changes from the
Pilot Survey Report. After publication of the Pilot Survey Report, revisions were made on two variables
in the Pilot Survey data, A1 and ALKA11. The revised data for the Southern Blue Ridge are included
only in Data Set NSSIDS4.
REVISIONS TO A1 (a,)
It is important to note any change to A1, because this variable is used in calculating the final
weight of each reach (the number of reaches the observation represents in the target population).
During the final assessment of NSS geographic data, the A1 estimates for two reaches were revised.
The first was stream reach "2A07891,' a special interest site. The A1 value for this reach was revised
from 5.15 mi2 to 5.19 mi2. Since special interest sites are not used in making statistical estimates of
the target population, this update does not affect any estimates of the Southern Blue Ridge reported
by Messer et al. (1986). The second A1 revision, however, was for stream reach '2A07881," a proba-
bility sample site. The A1 value was revised from 13.34 mi2 to 6.57 mi2. In turn, this reach's final
sample weight changed by about 10, from 9.60 to 19.48, resulting in a change in the estimated target
population for the Southern Blue Ridge from 2,021 reaches to 2,031 reaches. The standard error
estimate also changed from 326.7 to 326.4. This change is minor in terms of the total estimate and
does not change any of the conclusions based on the results of the Pilot Survey.
REVISIONS TO ALKALINITY (ALKA11)
During the full-scale NSS-I, the method used to calculate ANC and BNC from Gran titration data
was refined. All ANC values for streams in the Southern Blue Ridge were revised. The revisions are
summarized as follows. All ANC values for ALKA11 are in yeq L"1.
Mean
ANC Standard
Change Deviation
Mean % Standard
Change Deviation
Mean Mean
Absolute Standard Absolute Standard
Change Deviation % Change Deviation
1.85
31.22
0.9%
6.6%
7.3
30.4
3.0%
5.9%
76
-------
The relative range of revision to ANC values was generally within 7 (ieq L"1 of the original value
(Figure C-1). The revised ANC values are, on the whole, slightly greater than the original values.
Regression of new values on original values shows the magnitude of difference, with a slope of 1.017
and an r2 of 0.9908 (Figure C-2a). This small adjustment is minimized even further when these data
are considered in the context of the target population estimates, which are normally based on index
values, each of which is the mean of three spring measurements (Figure C-2b), with a regression
slope of 1.008 and an r2 of 0.9964. Overall, the revision in ANC calculation had minimal impact on
estimates of the distribution of ANC in the target population. Figure C-3 compares the cumulative
distribution function (CDF) of the target population for the original Pilot Survey ANC values and the
revised values.
77
-------
30-f
0 100 200 300 400 500
-30
2000
Figure C-1. Difference in original and revised ANC values versus original ANC.
78
-------
2000-r
1500 J
>
o
1000 J
1
500 4
500 1000 1500
REVISED ANC VALUES
2000
2000 H
1500 4
£3
I
o
1000 4
a:
O
500 J
500 1000 1500
REVISED ANC VALUES
2000
Figure C-2. NSS-I Pilot Survey original versus revised ANC values for (a) all samples, and (b) NSS-I
Pilot Survey original versus revised ANC values using averaged "indexed" observations.
79
-------
1.0-
0.9-
Z0.8-
O
£0.71
§0.61
cc
0.
111 0.5-
g
"~ 0.4-
0.3-
O
0.2-
0.1
0.01
NUMBER OF REACHES
-------
APPENDIX D
NSS-I FIELD OBSERVATION VARIABLES
Data Set NSSFSO
FIELD SITE OBSERVATIONS
Originally categorized as 'Watershed Characteristics', this information pertains only to the area in
the immediate vicinity of the sampling site. Table D-1 is a list of the variables. The field site obser-
vation data contain four categories of information:
• Descriptive Information
• Watershed Activities/Disturbances
• Bank Coverage
• Stream substrate
Descriptive Information
This includes basic identification variables and location specific information about the sampling
site.
Watershed Activities/Disturbances
Field crews recorded observations of immediate watershed characteristics but did not perform an
extensive field reconnaissance of the entire watershed area for disturbances which may influence
reach chemistry. These data were collected to document potential watershed influences on sample
chemistry and do not identify all potential disturbances in the entire watershed. A standardized format
was used to record observations about potential watershed impacts and human disturbances. The
approximate distance of disturbance from a sampling location was estimated to the nearest 100 feet
(30 meters). Table D-2 summarizes watershed activity and disturbance variables. Although none of
this information has undergone the level of quality assurance that was applied to the chemical data,
such information is helpful when interpreting individual sample chemistries.
Bank Coverage
Bank coverage estimates were made for the area within 100 meters of the stream bed. This
category provides an estimate of vegetation cover and type based on a coarse scale of high, medium,
low, and none.
81
-------
Stream Substrate
Substrate composition estimates were made at the reach sampling location. Particle size
categories were based on a scheme suggested by Cummins (1962). This information is based on
observations of individual field crews and is on the whole subjective. This information has not been
subjected to the stringent quality assurance review that the chemical data have. Observations of
substrate are difficult to review, but these data may be useful for examining the specific conditions that
existed at the time of sampling.
82
-------
TABLE D-1. NSS-I FIELD SITE OBSERVATIONS
SAMPLING SITE IDENTIFICATION DATA
Parameter
Stream ID code
Latitude(ddmmss)
Stream depth (m)
Stream width (m)
Field Comments
Variable
Name
STRM ID
LAT ST
STRMDP
STRMWD
COMM07
STREAM
Parameter
Stream name
Longitude(dddmmss)
1 :250,000 Map name
1 :24,000 Map name
Variable
Name
STRMNAM
LON ST
MAP BIG
MAP SML
Parameter
Elevation(ft)
Date Sampled
Sample Crew ID
County name
Variable
Name
ELEV
DATSMP
CRW ID
COUNTY
BANK COVERAGE ESTIMATES WITHIN 100 METERS OF STREAM BED:
A-Absent; S-Sparse (<25%);
Parameter
%Conif. trees
%Rock/bare
%Wetland
Variable
Name
CNTREE
RCKBR
WETLND
Parameter
%Deciduous trees
Shrub cover
M-Moderate (25%-75%)
Variable
Name
OCTREE
SHRUB
STREAM SUBSTRATE COVERAGE ESTIMATES WITHIN 100
A-Absent; S-Sparse (< 25%);
Parameter
%Substrate:aufwuchs
%Substrate:gravel
Variable
Name
AUFS
GRAVEL
Parameter
M-Moderate (25%-75%)
Variable
Name
%Substrate:boulders BOULD
%Substrate:sand SAND
; H-Heavy (>75%)
Parameter
Grasses/forb
%Moss cover
METERS OF STREAM BED:
; H-Heavy (> 75%)
Parameter
%Substrate:cobble
%Substrate:silt
Variable
Name
GRASS
MOSS
Variable
Name
COBBLE
SILT
-------
TABLE D-2. WATERSHED ACTIVITY/DISTURBANCE VARIABLE SUMMARY
Parameter
Variable
Name
Parameter
Variable
Name
Roadways Along Stream
Distance to unpaved road (m)
Crossings Above Stream
Distance to bridge (m)
Distance to grade (m)
Dwellings
Distance to multiple dwellings (m)
Agriculture
Distance to cropland (m)
Distance to unfenced land (m)
5° Industry
Industry! (type), Distance (m)
Logging
Presence and Age, Distance (m)
Fires
Presence and Age, Distance (m)
Mines/Quarries
Type, Distance (m)
Impoundments
Above site (type), Distance (m)
Livestock
Livestock (type), Distance from (m)
Other
Other disturbances near site, Distance (m)
UPRD_D
BRDG_D
GRAD_D
MDWL_D
CROP_D
UFNC_D
IND1, IND1_D
LOG, LOG_D
FIRE, FIRE_D
MNQR, MNQR_D
IMPA, IMPA_D
LIVE, LIVE_D
OTH, OTH D
Distance to paved road (m)
Distance to culvert (m)
Distance to single dwellings (m)
Distance to pastures (m)
Distance to fenced land (m)
Below site (type, Distance (m)
PRD_D
CULV_D
SDWL_D
PSTR_D
FENC D
IMPB, IMPB_D
-------
APPENDIX E
CARD IMAGE FORMAT DEFINITION
Only the final NSS-I data sets (Data Set 4 for the Pilot Survey and the full-scale NSS-I Survey) are
provided as both SAS-formatted files and as 80-column ASCII card-image files. This includes data
sets NSSIDS4, SBRSYN, and PILOTDS4. The data for NSSIDS4 have been divided into two data sets
(NSSIDS4A and NSSIDS4B), so that each will fit on a 1.2 M floppy diskette. Data Set NSSIDS4A con-
tains data for streams in subregions 1D, 2Bn, 2Cn, and 3B. Data Set NSSIDS4B contains data for
streams in subregions 2As, 2D, 2X, 3A, and 3C, and for all special interest sites. The formats for all
four data sets are listed in Tables F-1, F-2, and F3. Numeric variables were transferred to the card-
image files using the suggested variable width listed in Table B-1. Dates are in DDMMMYY format and
times are in HH:MM format (24-h clock) for all data card-image data sets. Missing numeric variables
are represented as -999. These values should be removed prior to any data analysis.
85
-------
TABLE E-1. Card-image Format Definition, NSS Data Sets NSSIDS4A and NSSIDS4B
Variable
Card # Name
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
5
6
6
6
6
6
6
6
6
7
7
7
7
7
8
8
9
9
10
10
10
11
11
11
11
11
11
12
12
12
12
12
12
12
12
12
12
13
13
A1
A1PRIME
A2
A3
A4
A5
ACC01 1
ALDS16
ALEX16
ALINOR
ALKA11
ALOR16
ALTL16
ANDEF
ANSUM
A US
CA16
CATSUM
CL16
C0316
COLVAL
COND11
CON IS
CONVAL
COUNTY1
COUNTY2
COUNTY3
COUNTY4
DATSMP
DICE11
DICI11
DICVAL
DOC 11
DO IS
DRPCDE
ELEV
FE16
FTL16
GRADE
H16
HC0316
K16
L2
LABNAM
LAT STD
LON STD
MART
HAP2
HAP3
MAP4
MAPS
MAP6
MG16
MN16
NA16
NH416
N0316
NODE
NOTSAM
OH16
ORGION
PHAC11
PHAL11
PHEQ11
PHSTVL
PH CLO
PH R
PTD16
PTL16
QUAD
RCH HU
Variable
Type
Num
Nun
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Char
Char
Char
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Num
Num
Char
Char
Char
Char
Char
Char
Num
Num
Num
Num
Num
Char
Char
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Num
Format
7.3
7.3
7.3
7.3
7.3
7.3
9.3
9.3
9.3
9.3
9.3
9.3
9.3
9.3
9.3
7.3
9.3
9.3
9.3
9.3
8.0
9.3
9.3
9.3
15.0
15.0
15.0
15.0
7.0
9.3
9.3
9.3
9.3
9.3
2.0
7.2
9.3
9.3
8.2
9.4
9.3
9.3
9.3
6.0
9.4
9.4
32.0
32.0
32.0
32.0
32.0
32.0
9.3
9.3
9.3
9.3
9.3
1.0
30.0
9.3
9.3
4.2
4.2
4.2
4.2
4.2
4.2
9.3
9.3
30.0
2.0
Column
Start
0
9
18
27
36
45
54
65
0
11
22
33
44
55
66
0
9
20
31
42
53
63
0
11
22
39
56
0
17
26
37
48
59
0
11
15
24
35
46
56
67
0
11
22
30
41
0
34
0
34
0
34
68
0
11
22
33
44
47
0
11
22
28
34
40
46
52
58
69
0
32
Column
End
7
16
25
34
43
52
63
74
9
20
31
42
53
64
75
7
18
29
40
51
61
72
9
20
37
54
71
15
24
35
46
57
68
9
13
22
33
44
54
65
76
9
20
28
39
50
32
66
32
66
32
66
77
9
20
31
42
45
77
9
20
26
32
38
44
50
56
67
78
30
34
Label Card #
DIRECT WATERSHED AREA (SQ MI)
UPDATED (1989) A1 (SQ MI)
US AREA TO MAPPED UPPER NODE (SQ MI)
US AREA TO MAPPED HEADUATER (SQ MI)
US AREA BETUEEN U/L SAMPLE SITE (SQ KM)
US AREA TO UPPER SAMPLE SITE (SQ KM)
BASE NEUTRALIZING CAPACITY (UEQ/L)
MONOMERIC (PCV) ALUMINUM (UMOL/L)
EXTRACTABLE (MIBK) ALUMINUM (UMOL/L)
INORG. MONOMERIC ALUMINUM (UMOL/L)
ACID NEUTRALIZING CAPACITY (UEQ/L)
ORG. MONOMERIC (PCV) ALUMINUM (UMOL/L)
TOTAL ALUMINUM (UMOL/L)
ANION DEFICIT, CATSUM-ANSUM (UEQ/L)
SUM OF AN IONS (UEQ/L)
US AREA TO MAPPED NODE (SQ KM)
CALCIUM (UEQ/L)
SUM OF CATIONS (UEQ/L)
CHLORIDE (UEQ/L)
CARBONATE (UEQ/L)
COLOR VALUE (PCU)
CONDUCTANCE -ANALYTICAL LAB- (US/CM)
IN-SITU CONDUCTANCE (US/CM)
CONDUCTANCE -PROCESS. LAB- (US/CM)
COUNTY NAME
COUNTY NAME
COUNTY NAME
COUNTY NAME
DATE SAMPLED
AIR EQUIL. DIS. INORG. CARBON (MG/L)
INITIAL DIS. INORGANIC CARBON (MG/L)
DIS. INORG. CARBON -PROCESS. LAB- (MG/L)
DIS. ORGANIC CARBON (MG/L)
IN-SITU DISSOLVED OXYGEN (MG/L)
SITE EXCLUSION CODE (0,1,2,3,4,5,13)
SAMPLE SITE ELEVATION (M)
IRON (UMOL/L)
TOTAL FLUORIDE (UEQ/L)
STREAM REACH GRADIENT (%)
HYDROGEN ION ACTIVITY (UEQ/L)
BICARBONATE (UEQ/L)
POTASSIUM (UEQ/L)
LENGTH BETUEEN U/L SAMPLE SITES (KM)
CHEMICAL ANALYSIS LABORATORY NAME
SAMPLE SITE LATITUDE (DECIMAL FORM)
SAMPLE SITE LONGITUDE (DECIMAL FORM)
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
MAGNESIUM (UEQ/L)
MANGANESE (UMOL/L)
SODIUM (UEQ/L)
AMMONIUM (UEQ/L)
NITRATE (UEQ/L)
REACH SAMPLE POSITION (U=UPPER,L=LOUER)
REASON NOT SAMPLED
HYDROXIDE (UEQ/L)
CALCULATED ORGANIC ANIONS (UEQ/L)
INITIAL PH, ACIDITY TITRATION
INITIAL PH, ALKALINITY TITRATION
AIR EQUILIBRATED LAB PH
CLOSED SYSTEM PH -PROCESS. LAB-
FIELD PH, CLOSED CONTAINER -PILOT ONLY
FIELD PH, OPEN SYSTEM
TOTAL DISSOLVED PHOSPHOROUS (UMOL/L)
TOTAL PHOSPHOROUS (UMOL/L)
1:250,000 SCALE MAP NAME
SHREVE ORDER -1:250,000 SCALE MAP
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
3
4
4
4
4
4
5
5
5
5
5
5
6
6
6
6
6
6
6
6
7
7
7
7
7
8
8
9
9
10
10
10
11
11
11
11
11
11
12
12
12
12
12
12
12
12
12
12
13
13
86
-------
TABLE E-1. Card-image Format Definition, NSS Data Sets NSSIDS4A and NSSIDS4B (continued)
Card #
13
13
13
13
13
14
14
14
14
14
14
14
14
14
14
15
15
15
15
15
15
16
16
16
Variable
Name
RCH ID
RCH~LN
SAMCOD
SAMRN
SHRE75
SI0216
SIT CIS
S04T6
SOBC
STATE 1
STATE2
STRA75
STRA250
STRATUM
STRMDP
STRMNAM
STRMWD
STRM ID
SUB ID
TIHSMP
TMPSTR
TURVAL
U
UC
Variable
Type
Char
Num
Char
Num
Num
Num
Char
Num
Num
Char
Char
Num
Num
Num
Num
Char
Num
Char
Char
Num
Num
Num
Num
Num
Format
8.0
9.3
3.0
1.0
3.0
9.3
6.0
9.3
9.3
2.0
2.0
9.3
9.3
1.0
3.1
30.0
4.1
9.0
3.0
5.0
9.3
9.3
12.6
12.6
Column
Start
36
46
57
62
65
0
11
19
30
41
45
49
60
71
74
0
32
38
49
54
61
0
11
25
Column
End
44
55
60
63
68
9
17
28
39
43
47
58
69
72
77
30
36
47
52
59
70
9
23
37
Label
REACH IDENTIFICATION CODE
LENGTH OF MAPPED BLUE LINE REACH (KM)
SAMPLE TYPE (D,DA,E,EDA,ER,NS,SY,R)
SAMPLE VISIT NUMBER (0,1,2,3,4)
SHREVE ORDER -1:24,000 SCALE MAP
DISSOLVED SILICA (UMOL/L)
SITE CHARACTERISTIC CODE
SULFATE (UEQ/L)
SUM OF BASE CATIONS (UEQ/L)
STATE (TWO CHARACTER CODE)
STATE (TWO CHARACTER CODE)
STRAHLER ORDER -1:24,000 SCALE MAP
STRAHLER ORDER -1:250,000 SCALE MAP
STRATUM (1=REG.,2=LOW ANC,3=SMALL A1)
STREAM DEPTH (M)
STREAM NAME
STREAM WIDTH (M)
STREAM/SITE IDENTIFICATION CODE
SUBREGION IDENTIFICATION CODE
TIME SAMPLED (HH:MM)
STREAM TEMPERATURE (DEC C)
TURBIDITY (NTU)
REACH WEIGHTING FACTOR
STAGE II CONDITIONAL WEIGHT
Card #
13
13
13
13
13
14
14
14
14
14
14
14
14
14
14
15
15
15
15
15
15
16
16
16
87
-------
TABLE E-2. Card-image Format Definition, NSS Data Set SBRSYN
Card #
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
3
4
4
4
4
4
4
4
5
5
5
6
6
7
7
7
7
7
8
8
8
8
8
8
8
9
9
9
9
9
9
9
9
9
9
9
10
10
10
10
10
11
Variable
Name
A1
A1PRIME
A2
A3
A4
A5
ALEX16
ALKA11
ALOR16
ANDEF
ANSUM
A US
CA16
CAT SUM
CL16
COND11
COUNTY1
DICI11
DOC11
DRPCDE
ELEV
FE16
FTL16
GRADE
H16
HC0316
K16
L2
LAT STD
LON STD
MAPI
MAP2
MAP3
MAP4
MAPS
MG16
MN16
NA16
NH416
N0316
NODE
PHSTVL
PTL16
QUAD
RCH HW
RCH ID
RCH LN
SAMCOD
SAMRN
SHRE75
SI0216
S0416
SOBC
STATE1
STRA75
STRATUM
STRMDP
STRMNAM
STRMWD
STRM ID
SUB ID
U
we
Variable
Type Format
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Char
Char
Char
Char
Num
Num
Num
Num
Num
Char
Num
Num
Char
Num
Char
Num
Char
Num
Num
Num
Num
Num
Char
Num
Num
Num
Char
Num
Char
Char
Num
Num
7.3
7.3
7.3
7.3
7.3
7.3
9.3
9.3
9.3
9.3
9.3
7.3
9.3
9.3
9.3
9.3
15.0
9.3
9.3
2.0
7.2
9.3
9.3
8.2
9.4
9.3
9.3
9.3
9.4
9.4
32.0
32.0
32.0
32.0
32.0
9.3
9.3
9.3
9.3
9.3
1.0
4.2
9.3
30.0
2.0
8.0
9.3
3.0
1.0
3.0
9.3
9.3
9.3
2.0
9.3
1.0
3.1
30.0
4.1
9.0
3.0
12.6
12.6
Column
Start
0
9
18
27
36
45
54
65
0
11
22
33
42
53
64
0
11
28
39
50
54
63
0
11
21
32
43
54
65
0
11
45
0
34
0
34
45
56
67
0
11
14
20
31
63
67
0
11
16
19
24
35
46
57
61
72
75
0
32
38
49
54
0
Column
End Label Card #
7
16
25
34
43
52
63
74
9
20
31
40
51
62
73
9
26
37
48
52
61
72
9
19
30
41
52
63
74
9
43
77
32
66
32
43
54
65
76
9
12
18
29
61
65
75
9
14
17
22
33
44
55
59
70
73
78
30
36
47
52
66
12
DIRECT WATERSHED AREA (SQ MI)
UPDATED (1989) A1 (SQ MI)
WS AREA TO MAPPED UPPER NODE (SQ MI)
WS AREA TO MAPPED HEADWATER (SQ MI)
WS AREA BETWEEN U/L SAMPLE SITE (SQ KM)
WS AREA TO UPPER SAMPLE SITE (SQ KM)
EXTRACTABLE (MIBK) ALUMINUM (UMOL/L)
ACID NEUTRALIZING CAPACITY (UEQ/L)
ORG. MONOMERIC (PCV) ALUMINUM (UMOL/L)
ANION DEFICIT, CATSUM-ANSUM (UEQ/L)
SUM OF ANIONS (UEQ/L)
WS AREA TO MAPPED NODE (SQ KM)
CALCIUM (UEQ/L)
SUM OF CATIONS (UEQ/L)
CHLORIDE (UEQ/L)
CONDUCTANCE -ANALYTICAL LAB- (US/CM)
COUNTY NAME
INITIAL DIS. INORGANIC CARBON (MG/L)
DIS. ORGANIC CARBON (MG/L)
SITE EXCLUSION CODE (0,1,2,3,4,5,13)
SAMPLE SITE ELEVATION (M)
IRON (UMOL/L)
TOTAL FLUORIDE (UEQ/L)
STREAM REACH GRADIENT (%)
HYDROGEN ION ACTIVITY (UEQ/L)
BICARBONATE (UEQ/L)
POTASSIUM (UEQ/L)
LENGTH BETWEEN U/L SAMPLE SITES (KM)
SAMPLE SITE LATITUDE (DECIMAL FORM)
SAMPLE SITE LONGITUDE (DECIMAL FORM)
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
MAGNESIUM (UEQ/L)
MANGANESE (UMOL/L)
SODIUM (UEQ/L)
AMMONIUM (UEQ/L)
NITRATE (UEQ/L)
REACH SAMPLE POSITION (U=UPPER,L=LOWER)
CLOSED SYSTEM PH -PROCESS. LAB-
TOTAL PHOSPHOROUS (UMOL/L)
1:250,000 SCALE MAP NAME
SHREVE ORDER -1:250,000 SCALE MAP
REACH IDENTIFICATION CODE
LENGTH OF MAPPED BLUE LINE REACH (KM)
SAMPLE TYPE (D,DA,E,EDA,ER,NS,SY,R)
SAMPLE VISIT NUMBER (0,1,2,3,4)
SHREVE ORDER -1:24,000 SCALE MAP
DISSOLVED SILICA (UMOL/L)
SULFATE (UEQ/L)
SUM OF BASE CATIONS (UEQ/L)
STATE (TWO CHARACTER CODE)
STRAHLER ORDER -1:24,000 SCALE MAP
STRATUM (1=REG.,2=LOW ANC,3=SMALL A1 )
STREAM DEPTH (M)
STREAM NAME
STREAM WIDTH (M)
STREAM/SITE IDENTIFICATION CODE
SUBREGION IDENTIFICATION CODE
REACH WEIGHTING FACTOR
STAGE II CONDTIONAL WEIGHT
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
3
4
4
4
4
4
4
4
5
5
5
6
6
7
7
7
7
7
8
8
8
8
8
8
8
9
9
9
9
9
9
9
9
9
9
9
10
10
10
10
10
11
88
-------
TABLE E-3. Card-image Format Definition, NSS Data Det PILOTDS4
Card #
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
4
4
4
4
4
4
5
5
5
5
5
5
5
5
6
6
6
6
6
7
7
8
8
9
9
9
9
9
10
10
10
10
10
10
10
11
11
11
11
11
11
11
12
12
12
12
Variable
Name
A1
A1PRIME
A2
A3
A4
A5
ACC011
ALEX16
ALKA11
ALOR16
ALTL16
ANDEF
ANSUM
A US
CA16
CAT SUM
CL16
C0316
COLVAL
COND11
CON IS
COUNTY 1
DATSMP
DICE11
DICI11
DICVAL
DOC 11
DO IS
DRPCDE
ELEV
FE16
FTL16
GRADE
H16
HC0316
K16
L2
LABNAM
LAT STD
LOW STD
MAPI
MAP2
MAP3
MAP4
MAPS
MG16
MN16
NA16
NH416
N0316
NODE
NOTSAM
OH16
ORGION
PHAC11
PHAL11
PHEQ11
PHSTVL
PH CLO
PH R
PTL16
QUAD
RCH HU
RCH ID
RCH LN
SAMCOD
SAMRN
Variable
Type
Num
Nun
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Num
Num
Char
Char
Char
Char
Char
Num
Num
Num
Num
Num
Char
Char
Num
Num
Num
Num
Num
Num
Num
Num
Num
Char
Num
Char
Num
Char
Num
Format
7.3
7.3
7.3
7.3
7.3
7.3
9.3
9.3
9.3
9.3
9.3
9.3
9.3
7.3
9.3
9.3
9.3
9.3
8.0
9.3
9.3
15.0
7.0
9.3
9.3
9.3
9.3
9.3
2.0
7.2
9.3
9.3
8.2
9.4
9.3
9.3
9.3
6.0
9.4
9.4
32.0
32.0
32.0
32.0
32.0
9.3
9.3
9.3
9.3
9.3
1.0
30.0
9.3
9.3
4.2
4.2
4.2
4.2
4.2
4.2
9.3
30.0
2.0
8.0
9.3
3.0
1.0
Column
Start
0
9
18
27
36
45
54
65
0
11
22
33
44
55
64
0
11
22
33
43
54
0
17
26
37
48
59
0
11
15
24
35
46
56
67
0
11
22
30
41
0
34
0
34
0
34
45
56
67
0
11
14
46
57
68
74
0
6
12
18
24
35
67
0
10
21
26
Column
End
7
16
25
34
43
52
63
74
9
20
31
42
53
62
73
9
20
31
41
52
63
15
24
35
46
57
68
9
13
22
33
44
54
65
76
9
20
28
39
50
32
66
32
66
32
43
54
65
76
9
12
44
55
66
72
78
4
10
16
22
33
65
69
8
19
24
27
Label
DIRECT WATERSHED AREA (SQ MI)
UPDATED (1989) A1 (SQ MI)
US AREA TO MAPPED UPPER NODE (SQ MI)
US AREA TO MAPPED HEADUATER (SQ MI)
US AREA BETUEEN U/L SAMPLE SITE (SQ KM)
US AREA TO UPPER SAMPLE SITE (SQ KM)
BASE NEUTRALIZING CAPACITY (UEQ/L)
EXTRACTABLE (MIBK) ALUMINUM (UMOL/L)
ACID NEUTRALIZING CAPACITY (UEQ/L)
ORG. MONOMERIC (PCV) ALUMINUM (UMOL/L)
TOTAL ALUMINUM (UMOL/L)
ANION DEFICIT, CATSUM-ANSUM (UEQ/L)
SUM OF ANIONS (UEQ/L)
US AREA TO MAPPED NODE (SQ KM)
CALCIUM (UEQ/L)
SUM OF CATIONS (UEQ/L)
CHLORIDE (UEQ/L)
CARBONATE (UEQ/L)
COLOR VALUE (PCU)
CONDUCTANCE -ANALYTICAL LAB- (US/CM)
IN-SITU CONDUCTANCE (US/CM)
COUNTY NAME
DATE SAMPLED
AIR EQUIL. DIS. INORG. CARBON (MG/L)
INITIAL DIS. INORGANIC CARBON (MG/L)
DIS. INORG. CARBON -PROCESS. LAB- (MG/L)
DIS. ORGANIC CARBON (MG/L)
IN-SITU DISSOLVED OXYGEN (MG/L)
SITE EXCLUSION CODE (0,1,2,3,4,5,13)
SAMPLE SITE ELEVATION (M)
IRON (UMOL/L)
TOTAL FLUORIDE (UEQ/L)
STREAM REACH GRADIENT (%)
HYDROGEN ION ACTIVITY (UEQ/L)
BICARBONATE (UEQ/L)
POTASSIUM (UEQ/L)
LENGTH BETUEEN U/L SAMPLE SITES (KM)
CHEMICAL ANALYSIS LABORATORY NAME
SAMPLE SITE LATITUDE (DECIMAL FORM)
SAMPLE SITE LONGITUDE (DECIMAL FORM)
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
1:24,000 SCALE MAP NAME
MAGNESIUM (UEQ/L)
MANGANESE (UMOL/L)
SODIUM (UEQ/L)
AMMONIUM (UEQ/L)
NITRATE (UEQ/L)
REACH SAMPLE POSITION (U=UPPER,L=LOWER)
REASON NOT SAMPLED
HYDROXIDE (UEQ/L)
CALCULATED ORGANIC ANIONS (UEQ/L)
INITIAL PH, ACIDITY TITRATION
INITIAL PH, ALKALINITY TITRATION
AIR EQUILIBRATED LAB PH
CLOSED SYSTEM PH -PROCESS. LAB-
FIELD PH, CLOSED CONTAINER -PILOT ONLY
FIELD PH, OPEN SYSTEM
TOTAL PHOSPHOROUS (UMOL/L)
1:250,000 SCALE MAP NAME
SHREVE ORDER -1:250,000 SCALE MAP
REACH IDENTIFICATION CODE
LENGTH OF MAPPED BLUE LINE REACH (KM)
SAMPLE TYPE (D,DA,E,EDA,ER,NS,SY,R)
SAMPLE VISIT NUMBER (0,1,2,3,4)
Card #
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
3
3
3
3
3
3
4
4
4
4
4
4
5
5
5
5
5
5
5
5
6
6
6
6
6
7
7
8
8
9
9
9
9
9
10
10
10
10
10
10
10
11
11
11
11
11
11
11
12
12
12
12
89
-------
TABLE E-3. Card-image Format Definition, NSS Data Det PILOTDS4
Card #
12
12
12
12
12
12
13
13
13
13
13
13
13
13
14
14
14
14
14
Variable
Name
SHRE75
SI0216
SIT CIS
S0416
SOBC
STATE 1
STATE2
STRA75
STRATUM
STRHOP
STRMNAM
STRHUD
STRH ID
SUB ID
TIMSMP
TMPSTR
TURVAL
W
we
Variable
Type
Num
Hum
Char
Num
Num
Char
Char
Num
Num
Num
Char
Num
Char
Char
Num
Num
Num
Num
Num
Format
3.0
9.3
6.0
9.3
9.3
2.0
2.0
9.3
1.0
3.1
30.0
4.1
9.0
3.0
5.0
9.3
9.3
12.6
12.6
Column
Start
29
34
45
53
64
75
0
4
15
18
23
55
61
72
0
7
18
29
43
Column
End
32
43
51
62
73
77
2
13
16
21
53
59
70
75
5
16
27
41
55
Label
SHREVE ORDER -1:24,000 SCALE HAP
DISSOLVED SILICA (UMOL/L)
SITE CHARACTERISTIC CODE
SULFATE (UEQ/L)
SUM OF BASE CATIONS (UEQ/L)
STATE (TWO CHARACTER CODE)
STATE (TWO CHARACTER CODE)
STRAHLER ORDER -1:24,000 SCALE MAP
STRATUM (1=REG.,2=LOW ANC,3=SMALL A1)
STREAM DEPTH (M)
STREAM NAME
STREAM WIDTH (M)
STREAM/SITE IDENTIFICATION CODE
SUBREGION IDENTIFICATION CODE
TIME SAMPLED (HH:MM)
STREAM TEMPERATURE (DEC C)
TURBIDITY (NTU)
REACH WEIGHTING FACTOR
STAGE II CONDTIONAL WEIGHT
Card #
12
12
12
12
12
12
13
13
13
13
13
13
13
13
14
14
14
14
14
90
-------
TABLE E-4. Card-image Listing (First Five Observations), Data Set NSSIDS4A, U.S. EPA National Stream
Survey
012345678
012345678901234567890123456789012345678901234567890123456789012345678901234567890
0.980 0.980 21.250 0.000 2.564 54.903 43.900 0.619
0.185 0.011 740.400 0.608 1.594 13.679 1176.559
57.576 848.300 1190.238 166.157 5.373 5 125.000
87.000 125.200 CHAUTAUQUA
23APR86 9.220 9.370 8.943 1.870
11.200 0 410.55 0.179 2.316 0.61 0.0126 721.503
18.487 2.108 NYSDOH 42.2150 79.0992
KENNEDY, NY 1979 CHERRY CREEK, NY 1954
HAMLET, NY 1954 GERRY, NY 1979
205.650
0.055 116.580 1.209 43.067 L
0.794 18.531 7.58 7.56 7.75 7.90 -999 8.24 0.084 -999.000
BUFFALO 1962 4 1D022009 2.172 R 1 19
56.088 237.348 1189.017 NY 4.000 2.000 3 0.3
CLEAR CREEK 7.0 1D022009L 1D 12:25 10.500
1.200 65.306122 1.000000
0.980 0.980 21.250 0.000 2.564 54.903 36.200 0.334
0.089 0.000 1324.200 0.552 0.519 8.474 1789.925
57.570 1312.370 1798.399 237.528 8.316 10 186.000
118.000 187.600 CHAUTAUQUA
08MAY86 14.400 14.900 15.210 0.843
10.600 0 410.55 0.125 2.421 0.61 0.0138 1224.260
22.962 2.108 NYSDOH 42.2150 79.0992
KENNEDY, NY 1979 CHERRY CREEK, NY 1954
HAMLET, NY 1954 GERRY, NY 1979
315.878
0.055 146.160 1.015 52.261 L
0.724 8.370 7.73 7.73 8.29 7.86 -999 7.98 0.132 -999.000
BUFFALO 1962 4 1D022009 2.172 R 2 19
27.794 264.414 1797.370 NY 4.000 2.000 3 0.3
CLEAR CREEK 7.0 1D022009L 1D 7:05 8.700
0.370 65.306122 1.000000
0.980 0.980 21.250 0.000 2.564 54.903 42.000 0.678
0.222 0.000 763.100 0.778 1.371 38.376 1183.386
55.032 878.240 1221.762 162.772 3.706 10 126.400
89.000 121.400 CHAUTAUQUA
23APR86 9.100 9.050 9.231 1.610
10.900 0 423.35 0.215 2.421 0.61 0.0186 735.925
18.768 2.108 NYSDOH 42.2264 79.1156
KENNEDY, NY 1979 CHERRY CREEK, NY 1954
HAMLET, NY 1954 GERRY, NY 1979
213.053
0.091 110.490 1.192 44.841 U
0.537 15.965 7.63 7.65 7.84 7.73 -999 7.92 0.107 -999.000
BUFFALO 1962 4 1D022009 2.172 R 1 19
51.262 233.184 1220.552 NY 4.000 2.000 3 0.2
CLEAR CREEK 6.0 1D022009U 1D 13:25 10.500
0.800 65.306122 1.000000
0.980 0.980 21.250 0.000 2.564 54.903 54.100 0.359
0.170 0.000 1277.000 0.567 0.852 49.396 1770.054
55.032 1347.300 1819.450 226.808 4.356 10 180.200
113.000 182.200 CHAUTAUQUA
08MAY86 15.200 15.700 15.570 2.420
9.800 0 423.35 0.090 2.316 0.61 0.0263 1221.937
22.962 2.108 NYSDOH 42.2264 79.1156
KENNEDY, NY 1979 CHERRY CREEK, NY 1954
HAMLET, NY 1954 GERRY, NY 1979
306.007
0.036 142.245 0.909 49.842 U
0.380 24.030 7.74 7.73 8.23 7.58 -999 7.69 0.042 -999.000
BUFFALO 1962 4 1D022009 2.172 R 2 19
36.615 264.414 1818.514 NY 4.000 2.000 3 0.2
CLEAR CREEK 6.0 1D022009U 1D 6:15 9.800
0.350 65.306122 1.000000
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
012345678
012345678901234567890123456789012345678901234567890123456789012345678901234567890
91
-------
TABLE E-4. Card-image Listing (First Five Observations), Data Set NSSIDS4A, U.S. EPA National Stream
Survey (continued)
012345678
012345678901234567890123456789012345678901234567890123456789012345678901234567890
18.470 18.470 0.000 0.910 45.243 2.564 91.500 0.574 1
0.111 0.000 1020.900 0.949 3.373 29.261 1537.182 2
47.837 1087.820 1566.443 210.164 4.607 15 162.000 3
97.000 142.400 CATTARAUGUS 4
23APR86 12.500 12.900 12.330 3.070 5
11.800 0 386.47 0.609 2.790 1.06 0.0200 980.334 6
25.826 13.807 NYSDOH 42.1561 78.9656 7
RANDOLPH, NY 1979 NEW ALBION, NY 1963 8
9
282.152 10
0.255 169.650 0.976 53.552 L 11
0.501 30.284 7.32 7.36 8.05 7.70 -999 7.92 0.158 -999.000 12
BUFFALO 1962 1 1D022010 14.419 R 1 28 13
56.421 285.234 1565.447 NY 4.000 1.000 1 0.2 14
ELM CREEK 7.0 1D022010L 1D 9:30 6.500 15
1.300 59.738953 17.240288 16
012345678
012345678901234567890123456789012345678901234567890123456789012345678901234567890
92
-------
SUBREGIONS OF THE NATIONAL STREAM SURVEY-PHASE I
Northern
Appalachians (2Cn)
Valley and Ridge (2Bn)
Southern Blue Ridge (2As)
(Pilot Study)
Poconos/Catskills (ID)
NY\
Ozarks/Ouachitas (2D)
Mid-Atlantic
Coastal Plain (3B)
Southern Appalachians (2X)
------- |