What are Quality Assurance and Quality Control (QA/QC)?
QA/QC basically refers to all those things good investigators
do to make sure their measurements are right on (accurate;
the absolute true value), reproducible (precise; consistent),
and have a good estimate of their uncertainty. In the regulatory
arena, this aspect of data collection is as crucial to the
final outcome of a confrontation as the numbers themselves.
It specifically involves following established rules in the
field and lab to assure everyone that the sample is representative
of the site, free from outside contamination by the sample
collector (no dirty hands touching the water) and that it
has been analyzed following standard QA/QC methods. This typically
involves comparing the sample to a set of known samples (calibration
standards) for estimating accuracy and by replicating the
measurement to estimate its precision. The U.S. Environmental
Protection Agency and the U.S. Geological Survey are federal
agencies that have lots to add should you wish to learn more
of the technical aspects of a Quality Assurance Program (QAP)
and creating a Quality Assurance Project Plan (QAPP).
Proper Quality Assurance and Quality Control (QA/QC)
protocols are essential to the LakeSuperiorStreams project and therefore
we have gone to great lengths to try to assure the accuracy
of our data. Quality Assurance is a particularly important component
of any program that posts real-time or near-real time data on
a Website that is freely available to the public. Data for this
project will be collected primarily by LakeSuperiorStreams staff from
the Natural Resources Research Institute (NRRI) at the University
of Minnesota-Duluth (UMD) with additional help from the City
of Duluth Stormwater Utility, the Minnesota Pollution Control
Agency (MPCA), and the Western Lake Superior Sanitary District
The methods used for data collection, storage and retrieval
as well as quality control and data delivery will closely
follow those described in the EPA-EMPACT Program's technology
transfer manual based on the Lake Access - Minneapolis Project
(EPA 2000). This project was developed by our NRRI-UMD group
working in collaboration with limnology staff at the Hennepin
Parks - Three Rivers Park District in the Minneapolis, MN
Metropolitan area. Other aspects of the "flow" of
information from real-time "data hoses" are described
in Host et al (1999, 2000).
There are basically four sets of environmental data that are
collected for LakeSuperiorStreams:
- Water quality parameters analyzed in the lab from "grab"
samples of water such as nutrients (N- and P-series of nutrients),
pH and alkalinity, turbidity, total suspended solids (TSS),
total dissolved solids (TDS), color, chloride, fecal coliform
bacteria, total volatile solids (TVS) and biochemical oxygen
demand (TVS and BOD5 are both measures of organic matter);
- Water quality measurements made in the field at the
time of water sample collection such as sensor measurements
of temperature, water velocity (to estimate flow), dissolved
oxygen (DO), specific electrical conductivity (EC25) and transparency
tube measurements of water transparency or clarity, etc.; and
- Remotely sensed water quality measurements collected,
logged and transmitted by electronic sensors deployed in
the streams that we call SMU's for Stream Monitoring Units.
These instruments are programmed to record temperature,
stream elevation (also called stage height), EC25 (a measure
of total salt concentration), and turbidity (used as a measure
of suspended sediment). In addition, the SMU at the Kingsbury
Creek site is connected to an automated sampling device
(ISCO 6712) that upon preset increases in stage height,
signifying a storm event, initiates sample collection of
up to 24 individual water samples into one liter plastic
bottles. We have initially programmed the unit to sample
at 2 hour intervals during storm events.
The stream sensor data is typically collected every 15 minutes,
stored throughout the day and then transmitted via cellular
phone/modem to our base computer/website each morning. We
can collect the data more frequently, and call up the unit
to see what the latest data looks like, but this adds to our
phone bill. A final intensive data set is collected at the
Duluth-Superior Harbor/ St. Louis River outlet to Lake Superior
where a sensor array measures and logs temperature, EC25 and
turbidity at 15 minute intervals. This unit was installed
on the channel wall above the USGS Superior Bay Duluth Ship
Channel (Station 464646092052900) where an acoustical velocity
meter (AVM) system with a two-path transducer installation
is used to measure river discharge into Lake Superior. We
have previously established a data-sharing partnership with
USGS (Madison, WI, P. Hughes) to link our on-line water quality
data with their discharge data.
- Volunteer monitoring data. During this grant period,
LakeSuperiorStreams staff will work with the MPCA to coordinate
existing volunteer monitoring with the goal of achieving
more consistent methodology and developing a repository
and database for stream-related information in the City
of Duluth. Programs include the state-wide Citizens Stream
Monitoring Program (MPCA sponsored), St. Louis River Watch
and a variety of less formal efforts by local schools and
environmental learning centers. LakeSuperiorStreams initiated
school-based stream monitoring programs at sites on Chester,
Tischer and Kingsbury Creeks in October 2002 as part of
the National Monitoring Day, nation-wide celebration
of the 30th anniversary of the Clean Water Act on October
15. It was subsequently decided to include these schools
in the St. Louis River Watch program which already includes
over 30 area schools (further information).
Parameters will vary depending on the level of effort available,
but at a minimum will include temperature, transparency tube
depth, turbidity and EC25 during seasonal baseflow conditions,
snowmelt runoff and rainstorm runoff. Training will be provided
by River Watch and LakeSuperiorStreams staff, and turbidity and
EC25 measurements will be performed by NRRI on discrete samples
collected by the schools. Biological monitoring of macroinvertebrate
communities is fundamental to River Watch, and LakeSuperiorStreams
will work to encourage standardized sampling protocols that
conform to peer-reviewed research standards and existing state
and federal guidelines. The goals of this effort are to initiate
a comprehensive database, standardize monitoring methods,
and educate students and citizens.
Conventional data quality assurance procedures follow guidelines
set by the U.S. EPA (1989a,b), and the so-called "Standard
Methods" from APHA (1998). Water chemistry and manual
field profiles are collected by trained staff aquatic ecologists
and technicians at NRRI-UMD (supervised by Co-Principal Investigator
/ Limnologist Rich Axler) and Minnesota Sea Grant-UMD (supervised
by Co-Principal Investigator / Aquatic Ecologist Carl Richards).
City of Duluth Stormwater Utility technicians have worked
closely with UMD staff to ensure that consistent sampling
and SMU maintenance procedures have been used. All water chemistry
analyses are performed by either the NRRI Central Analytical
Laboratory or the WLSSD laboratory which are certified annually
by the Minnesota Department of Health for Federal Safe Drinking
Water Act and Clean Water Act parameters. The certification
procedure involves blind analyses of certified performance
standards and an in-depth site inspection and interview approximately
every other year. The NRRI lab has also been certified over
the past decade by the Minnesota Pollution Control Agency
and the Minnesota Department of Natural Resources for low-level
water quality analyses in pristine, acid-sensitive lake monitoring
programs and for sediment contaminant analyses in the St.
Louis River and Upper Mississippi Rivers. Further details
regarding the water chemistry analyses are provided below.
SENSOR RESOLUTION & REPORTING LIMITS
Ideally, if all of the sensors behaved according to the sensor
manufacturer's specifications (Table 1) we could simply post
the data on the LakeSuperiorStreams web site and assume it is accurate
to these levels. However, except for temperature, all of the
sensors require routine maintenance and calibration. When
visiting sites we re-calibrate the EC and turbidity sensors
using individual standard solutions with known values for
EC25 and turbidity. Experience has taught us that the sensors
remain stable during the course of a sampling day. The depth
sensor is also calibrated at this time to read zero when the
sensor is set at the water surface.
When deployed for continuous operation the sensors are colonized
gradually by a biofilm of algae and less noticeably, by bacteria
and fungi as well. As this material builds up, the biofilm
interferes with the sensor's ability to accurately sample
the surrounding water. One can easily picture fine filaments
of algae, like you see on the rocks in summer, wafting intermittently
across the electrodes of the EC sensor or in the light path
of the turbidimeter. This would produce erratic values with
wide swings even during periods of low and relatively constant
flow when the stream is running clear. Turbidity probes are
most susceptible to these changes, followed by dissolved oxygen
(not measured automatically at this time) and EC. Fine particulate
sediment will also be trapped and gradually contribute to
erroneous readings if not cleaned on a regular basis. Turbidity
sensors are "notorious" for their maintenance problems
- the YSI 6136 sensor we are using is called a "self-wiping"
sensor that mechanically wipes the optical window used for
the measurement just before a reading is taken. Although this
feature certainly reduces errors due to fouling, the wiper
occasionally "sticks", creating apparent "spikes"
in the data set (see below). The Hydrolab self wiping turbidity sensors
have not experienced these wiper sticking issues as of yet.
Other sources of variability include the natural variations
that occur in a stream due to eddies and larger debris suddenly
breaking free, events less likely to be seen in the calmer
water in deeper stream pools or in lakes. Also, the sensor
is actually "sampling" only a tiny (millimeter scale)
patch of water for less than a second every 15 minutes. Table
2 compares the theoretical sensor resolution with our best
estimate of the true accuracy of the sensor readings, incorporating
all sources of known error.
|Table 1. Reported automated sensor specifications
YSI 6920 sonde
YSI 6136 turbidity sensor
||± 0.15 °C
||± 0.5% reading +1
The greater of:
± 5% of reading
or 2 NTU
||± 0.003 m
YSI 6820 sonde
as above; a 6820 sonde instead of a 6920 sonde is used
for cost and data logging considerations associated with
the automated water sampler
|stage height sensor
triggers ISCO 6712 automatic water sampler
|St. Louis River
||YSI 6820 sensors
as above (w/o stage height)
Apprise Technologies RepDAR unit
(see http://lakeaccess.org/QAQC.html for details)
Hydrolab MS5 sonde
Self cleaning turbidity sensor
Stage height is measured by the MPCA via an ultrasonic distance
sensor and flow calculated from USGS derived rating curve.
||± 0.10 °C
||± 1% reading +1
0.1, up to 400 NTU;
1.0, 400-3000 NTU
± 1%, up to 100 NTU;
± 3%, from 100-400 NTU;
± 5%, from 400-3000 NTU
Table 2. LakeSuperiorStreams Reporting limits for Stream Monitoring Unit
sensor data (YSI 6920/6820)
The resolution, i.e. the smallest reading displayed
for a particular parameter, is likely to be considerably
lower than the error associated with differences in
time, with sensor drift and calibration accuracy.
is reported by the SMU sensors)
Accuracy (what we really trust)
Stream Monitoring Unit (SMU) PLACEMENT, FLOW CALIBRATION , SAMPLE COLLECTION
We attempted to place the water quality probes at representative
measurement points in the stream cross-section with allowances
made for protecting the units from debris and sediment where
necessary. Probes are inspected, calibrated, and maintained
following the manufacturers recommendations in addition to
following USGS (2000a, 2000b) procedures. The exact location
of each unit within a watershed was chosen based upon considerations
of security, stability regarding anticipated road and bridge
construction activities, and on upstream land use characteristics.
Because of this, summary interpretations as to water quality
differences between streams must be done with caution since
such a direct comparison was not intended.
Stream discharge is determined as in Anderson et al (2000)
and USGS (2000b) with rating curves developed and subsequently
used to calibrate flow. Stream depth in Tischer, Chester and Kingsbury
is measured remotely
using a pressure transducer and discharge is determined using
rating curves based upon a set of cross-sectional and discrete
depth in-stream velocity measurements made with a Marsh-McBirney
stream velocity meter over a range of discharge conditions.
Stream depth at Amity and Poplar is measured by an MPCA owned/operated/maintained
ultrasonic distance sensor and discharge calculated from a USGS derived flow rating curve.
Discrete water samples are collected at Kingsbury Creek at
various points along the hydrograph using an ISCO 6712 slaved
to the YSI 6820 stream elevation sensor. Periodic intensive
manual collections throughout high water periods as well as
during baseflow periods are conducted for Chester, Tischer,
Amity and Poplar Creeks.
NRRI staff follow the Instrument Manuals for calibration
and maintenance procedures. Our staff also have extensive
(i.e. decades collectively) experience with these calibration
procedures and with their importance in interpreting field
data and distinguishing systematic errors associated with
deteriorating, or bio-fouled probes. Our Lake Access, EMPACT
project is a companion to an earlier NSF-funded Advanced
Technology Education project entitled Water on the Web
now in its fifth year, that involved establishment of Apprise
Technologies, Inc. robotic water quality sensing RUSS units
on five Minnesota lakes. Since 1998 we have gained extensive
experience in dealing with the problems associated with continuous
sensor deployment and the resultant protocols are included
also in the Lake Access and WOW websites.
The following protocols have been set up to minimize these
biofouling and instrument drift effects to quality assure
the LakeSuperiorStreams data:
- Clean and re-calibrate sensors frequently (about every
2 weeks) and perform manual measurements with an independent
instrument at the same time (temperature and EC25) or soon
after in the lab (turbidity). The depth sensor is also calibrated
at this time by re-setting it to equal zero depth when placed
just above the water surface.
- Compare independent manual measurements with near-simultaneous
SMU data prior to cleaning (re-calibration). This provides
assurance that the previous period of data is accurate.
We calculate test statistics for each parameter as:
RPD (relative % difference) = |Manual-SMU| x 100 , and
DELTA = |Manual SMU| for each parameter. They
PASS according to rules in Table 3.
|Table 3. Quality Assurance
Criteria for SMU Sensors
||< 0.2 °C
||< 5 µS/cm
||< 5 NTUs
If the data "passes," it is considered acceptable
for the previous period. If not, we examine it in the context
of our understanding of the instrumentation, the stream's prior
automated and "manual" water quality data, its watershed
and the recent weather pattern and then either delete it from
the database or allow it to be posted. We have to be careful
not to delete anomalous data that may simply reveal real dynamic
changes. The sheer volume of data has been taxing and we lack
the resources to always be as current as we would like. In the
interim, data are posted as provisional. Dates of calibrations
and manual data are both posted in the DATA section of the website
and are available within easily accessible Excel files. The
stream data visualization tool (DVT) is also helpful in rapidly
displaying the data in a variety of formats to help identify
anomalous data. Calibration "date flags" will be added
to the control panel of the DVT to allow the user to more easily
keep track of calibration dates as the data stream is being
Although not yet implemented (as of December 2002), we are
also exploring a data quality classification rating system
such as used by the US Geological Survey for certain continuous
water quality records. The table below is taken from USGS
(Wagner, R.J., Boulger, R.W., Jr., Oblinger, C.J., and Smith, B.A., 2006, Guidelines and standard procedures for continuous
water-quality monitors—Station operation, record computation, and data reporting: U.S. Geological Survey Techniques
and Methods 1–D3, 51 p. + 8 attachments; accessed April 10, 2006, at http://pubs.water.usgs.gov/tm1d3 - Table 18).
This USGS document also discusses the use of such data when
there are extensive periods without data as well as other
considerations that are beyond the scope of our present objectives.
Accuracy ratings of continuous water-quality
≤ less than or equal to; ±, plus or
minus value shown; °C, degree Celsius; >, greater
than; %, percent; mg/L, milligram per liter; pH unit,
standard pH unit
||≤ ±0.2 °C
||> ±0.2 - 0.5 °C
||> ±0.5 - 0.8 °C
||> ±0.8 °C
||> ±3 - 10%
||> ±10 to 15%
||> ±15 %
||≤ ±0.3 mg/L
||> ±0.3 - 0.5 mg/L
||> ±0.5 to 0.8 mg/L
||> ±0.8 mg/L
||≤ ±0.2 unit
||> ±0.2 - 0.5 unit
||> ±0.5 - 0.8 unit
||> ±0.8 unit
||> ±5 - 10%
||> ±10 - 15%
OTHER CONTINUOUS WATER MONITORING DATA
in collaboration with the South S. Louis County Soil and Water
Conservation District (SSLCSWCD) has also operated several
continuous stage height monitors on Miller Creek, and these
data will also be posted. A limited number of water quality
samples have also been collected at the St. Louis River/Lake
Superior entry site.
DATA TRANSMISSION AND INITIAL QA SCREENING
Each SMU is called daily and data that has been collected
since the last call is downloaded. These new data files are
stored on the base station computer as comma-delimited ASCII
text files. The Kingsbury station is polled twice daily, at
6 AM and 2 PM. The Chester and Tischer sites are scheduled
to be polled at 6AM each day because of cell phone costs.
The Conversion Process
A program (the data importer) is now launched. It reads any
data files that have been created or changed since the last
time it was run, and converts the data to the format used
by the report generating and data visualization programs.
The data importer examines the beginning of these files and
checks to make sure that the contents (e.g. site name, parameter
names and units) correspond to what is expected. If not, an
error message is generated and no further action is taken
with this file. This will catch errors that could occur if,
for example, a new parameter is being read by the SMU, but
the property files used by the data importer for that site
haven't been updated to handle the change.
Now, each data line is read and converted to the "DataTable"
format used for storage. The importing program can reject
specific data if it is outside of a pre-defined range for
the given parameter. Data are automatically rejected if outside
the following ranges:
||< -1 °C or > 35 °C.
||< 50 µS/cm or > 3000 µS/cm
(subject to review after we've experienced spring snowmelt)
||< - 5 & > 2000, with values from
- 5 to 0 set equal to 0
If it encounters a value outside of the range it will generate
an appropriate error message in the log file, and disregard
the value. This helps eliminate invalid readings that could
be caused by sensor drift or SMU hardware problems. After
all of the new data files have been imported the data importer
triggers the appropriate programs that will create and send
updated reports and data-visualization data files to the website.
FINAL DATA REVIEW & POSTING
At present (December 2002), funding limitations have precluded
adherence to a rigorous schedule for removing the provisional
label from LakeSuperiorStreams automated data. In part this is due
to the need to review ancillary water chemistry data before
making final decisions when the data is questionable. All
manually sampled water quality data posted on the LakeSuperiorStreams
however, have passed QA/QC prior to being posted, although
this typically takes from 30-60 days after collection.
Despite regular maintenance and calibration schedules, occasional
SMU data anomalies still occur. To date, they have generally
been associated with the turbidity sensor although there was
great difficulty during summer 2002 in getting the stage height
sensors to function properly. Additional difficulties were
caused by the Minnesota DNR modifying the Chester Creek stream
channel in the vicinity of our SMU, causing excessive turbidity
values and requiring a new empirical calibration of the stage
height-flow relationship. There have also been start up problems
associated with precipitation monitoring and the modem link
to Kingsbury Creek. All of these problems were resolved by
about September 1, 2002.
The most troublesome anomalies are those that occur within
the calibration window of time, are not flagged by our automated
screening tools and are not unreasonable values in terms of
the range of values previously measured or expected for the
hydrologic conditions at the time. These errors will not be
trivial to identify and will require careful examination in
a complete fluvial/limnological (stream/watershed/climate)
context by professionals. The process is adequately described
as Best Professional Judgement (BPJ). In some cases data will
need to be adjusted by calculating correction factors when
there is accurate calibration data spanning the period in
question and when the results estimated by interpolation are
consistent with the rest of the data set. In other cases,
it will be necessary to simply reject the data - omitting
it from the website. A log of data deletions will be maintained
on the website, and an e-mail announcement sent to all teachers
and researchers known to be using the site for educational
or research purposes related to curricula associated with
LakeSuperiorStreams and Water-on-the-Web
Data collected under
ice: When streams
are ice covered the depth readings and flow calculations
are subject to errors:
- The stream can be running under the ice without an
air gap, essentially
pressurizing the stream -- increasing the apparent
depth and velocity.
- The streams can be running on top of the
ice (probably unaccounted for flow).
- Anchor Ice may change the channel cross section, making
our rating curves unreliable.
- Ice dams below the site
can change flows.
- Bank ice can constrict the
All of these things can happen at
the same place over the course of time as well.
We have chosen to flag this data as 'collected under ice
cover'. It will still show relative changes, especially
over the short-term, but it should be
noted that the reported depths and flows
are not necessarily accurate.
WATER SAMPLE ANALYSES
Data from the monitoring programs will follow established
procedures of the NRRI Central Analytical Laboratory
(Ameel et al. 1998) and WLSSD (2000a,b). Both laboratories are
annually certified for Clean Water Act and Safe Drinking
Water Act water quality parameters by the Minnesota Department of
Sample Collection: Samples for micronutrients
([nitrate+nitrite]-nitrogen, ammonium-nitrogen, total-nitrogen
[TN], orthophosphate-phosphorus [OP], total phosphorus [TP],
alkalinity, chloride, BOD5, total suspended solids [TSS],
total volatile solids [TVS], and total dissolved solids [TDS]
and fecal coliform bacteria are collected exclusively by trained
personnel (City of Duluth, WLSSD, MPCA, NRRI Cental Analytical
Laboratory personnel), according to individual analysis requirements
using an ISCO 6712 automated sampler at the Kingsbury site.
Manual samples are collected at various times of the year
to characterize Chester and Tischer Creeks at their SMU sites.
Several intensive sampling surveys will be conducted to characterize
high flow storm events and the snowmelt runoff period in the
Spring. Grab samples for fecal coliform bacteria will be collected
using sterile technique (APHA 1998). All samples are transported
to the laboratory on ice in dark coolers on the day of collection,
within a few hours after collection. A field data sheet accompanies
each sample. Sample sites for the collection of duplicate
samples and preparation of field blanks are chosen at random.
Manual samples at Amity and Poplar river are being collected by the MPCA.
Sample Preservation: All samples are initially
preserved by refrigeration and when filtration is necessary
it is performed as soon as possible after return to the lab.
If delay in analysis is anticipated, additional preservation
will be incorporated, as provided in individual methods and
according to approved guidelines (e.g. acidification or freezing).
The container label is marked when preservative is added.
Sample Identification and Tracking: Sample
containers are labeled in the field with date and time of
collection, location of sample, and person collecting the
sample. Field data sheets include all of the above information,
and also conditions specific to the project and lists the
individual parameters for which analysis is requested. For
stream monitoring, and other sampling events for which field
analysis are performed (such as conductivity, dissolved oxygen,
pH, temperature), an additional sheet is used which lists
site, stream conditions and other pertinent information. Water
samples delivered to the laboratory are assigned a laboratory
identification code and entered into a computerized database
(Microsoft Access) along with location, collection date and
required analytes. The code is used to keep a running tally
of all samples received, and also to track analysis as they
Instrumentation: Laboratory instruments are
calibrated, operated, and maintained properly according to
manufacturer's recommendations and this is a requirement for
State Laboratory Certification. Instruments are calibrated
each time they are used, if applicable. A log is kept of each
instrument where maintenance and general comments on instrument
performance are documented. For spectroscopy procedures, the
calibration is performed with reagent blanks and standards
of a matrix closely matching the samples. A typical standard
curve consists of at least five concentrations. The linearity
of each standard curve is monitored and must have a regression
coefficient (r2) greater than 0.99. Additional quality assurance
procedures that are a part of every micro- and macronutrient
run include blanks, QCCS (quality control check standards),
and internal spikes (recoveries of 80-120% or 85-115% are
typically required depending on analyte).
Data Storage and Archiving: After analysis,
raw data from labsheets are entered into spreadsheets to calculate
concentrations and compute the relevant QC statistics. The
analyst checks the quality control data against laboratory
limits, and if the data meets the requirements, the data is
saved and printed in hard copy. Both the original data sheet
as well as the spreadsheet printout are saved. On the hard
copy, the analyst checks the data, shows the quality control
calculations, and initials the completed sheet. This data
is reviewed by the Laboratory QA officer who also initials
the sheet. At this point the data is considered valid and
reportable. Data from spreadsheets is imported electronically
into the water quality database on the LakeSuperiorStreams-NRRI
file server. The server is backed up daily to prevent data
Ameel, J., E. Ruzycki and R.P. Axler. 1998 (reviewed annually).
Analytical chemistry and quality assurance procedures for
natural water samples. 6th edition. Central Analytical Laboratory,
NRRI Tech. Rep. NRRI/TR98/03.
Anderson, J., T. Estabrooks, and J. McDonnel. 2000. Duluth
Metropolitan Area streams snowmelt runoff study. Minnesota
Pollution Control Agency, St. Paul, MN. March 2000.
APHA.1998. Standard methods for the examination of water and
wastewater. Amer. Publ. Health Assoc.
EPA. 2000. Delivering timely water quality information to
your community: The Lake Access-Minneapolis project. EPA/625/R-00/012,
September 2000, U. S. Environmental Protection Agency, Office
of Research and Development, Cincinnati, OH, 45268, USA.
EPA. 1998. Guidance for Quality Assurance Project Plans EPA
QA/G-5. EPA/600/R-98/018, Feb 1998 (http://www.epa.gov/quality1/qs-docs/g5-final.pdf ). U.S. EPA, Washington, D.C. 20460
EPA. 1996. The Volunteer Monitor's Guide to: Quality Assurance
Project Plans. EPA 841-B-96-003, Sep 1996, U.S. EPA, Office
of Wetlands, Washington, D.C. 20460, USA (http://www.epa.gov/region09/qa/pdfs/vol_qapp.pdf)
EPA 1989a. Preparing perfect project plans. US EPA Risk Reduction
Engineering Laboratory, Cincinnati, OH, EPA/600/9-89/087.
EPA.1989b. Handbook of methods for acid deposition studies-Field
operations for surface water chemistry. Environmental Protection
Agency, Washington, D.C.
Host, G., N. Will, R.Axler, C. Owen and B. Munson. 2000. Interactive
technologies for collecting and visualizing water quality
data, URISA Journal (http://wow.nrri.umn.edu/urisa)
Host, G.E. , B. H. Munson, R. P. Axler, C. A. Hagley, G. Merrick
and C. J. Owen. 1999. Water on the Web: Students monitoring
Minnesota rivers and lakes over the Internet. AWRA Spec. Ed.
(Dec., 1999). (refereed: http://www.awra.org/proceedings/www99/w74/index.htm
USGS. 2000a. Guidelines and standard procedures for continuous
water-quality monitors: Site selection, field operation, calibration,
record computation, and reporting. R.J. Wagner, H.C. Mattraw,
G.F. Fritz and B.A. Smith. U.S. Geological Survey Techniques
of Water-Resources Investigations Report 00-4252 (http://water.usgs.gov/pubs/wri/wri004252/).
U.S. Geological Survey, Reston, Virginia, USA.
USGS. 2000b. National field manual for the collection of water-quality
data. U.S. Geological Survey Techniques of Water-Resources
Investigations, book 9, chaps. A1-A9, 2 v., variously paged.
WLSSD. 2000a. Laboratory Procedures Manual. Revision 2, Oct
1998. Western Lake Superior Sanitary District, Duluth, MN. 55807.
WLSSD. 2000b. Laboratory Quality Assurance Manual. (Revision
2, May 1998 with annual minor revisions). Western Lake Superior
Sanitary District., Duluth, MN 55807.
September 16, 2008