EPA-600/4-80-052
(November 1980
THE DEVELOPMENT OF A DEPLOYABLE
WATER QUALITY MONITORING SYSTEM
by
C. A. Whitehurst
6. D. Whitehouse
Division of Engineering Research
College of Engineering
Louisiana State University
Baton Rouge, Louisiana 70803
Grant No. R806313010
Project Officer
Victor W. Lambou
Advanced Monitoring Systems Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89114
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, Nevada 89114
-------
EPA-600/4-80-052
November 1980
THE DEVELOPMENT OF A DEPLOYABLE
WATER QUALITY MONITORING SYSTEM
by
C. A. Whitehurst
G. D. Whitehouse
Division of Engineering Research
College of Engineering
Louisiana State University
Baton Rouge, Louisiana 70803
Grant No. R806313010
Project Officer
Victor W. Lambou
Advanced Monitoring Systems Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89114
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, Nevada 89114
-------
I -~--
DISCLAIMER
This report has been reviewed by the Environmental Monitoring Systems
Laboratory, U.S. EnvironrTEntal Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
SUMMARY
The purpose of this project was to evaluate the stabil ity and accuracy of
off-the-shelf water quality sensors used for extended periods of time. The
Hydrolab Surveyor Model .60, Martek Mark IV and Ocean Data Model lOlA probes
were used in the test. .
Based on an extensive literature survey and laboratory tests, biological
foul ing was deemed one of the most serious problems in null ifying sensor
capability. Several possible methods to reduce biofouling include:
l.
2.
3.
4.
5.
6.
7.
in situ ultrasonic cleaning of the probes
in situ generation of sodium hypochlorite from sea water
local application of heat
light-proof environment to suppress algae growth
generation of a local "kill z~ne" utilizing toxic ~ntifouling paints
periodic injection of a biocide .
incorporation of toxic compounds into dissolved oxygen membrane.
Of these seven potential measures to eliminate biofouling, the local "kill
zone" approach appears to be the most feasible and simplest to implement. In
field tests at Louisiana State University (LSU), it proved to be effective in
controlling biological growths.
An experimental program was designed to evaluate probe performance under
field conditions. Tests were conducted in a freshwater environment for
several extended periods during which calibration procedures and the stability
of calibration were studied. A statistical study of parameter variation is
included. .
Comparative results of the various probes are given for temperature,
dissolved oxygen (DO), pH and conductivity for a 6-week test period in a
freshwater environment.
iii
-------
.~-
CONTENTS
Page
Summa ry . . . . . . . . . . . . . . iii
Fi gures . . . . . . . . . . vi
Tables. . . . . . . . . . . vii
Abbreviations and Symbols . . . . . . viii
1. Introduction. . . . . . . . . . 1
2. Conclusions . . . . . . . 3
Temperature . . . . . . . . 4
pH. . . . . . . . . . . . . 4
Conductivity. . . . . 4
Dissolved oxygen. . . 5
Biofouling and "kill zone". . . . . 5
3. Recommendations . . . . . . 6
4. Background. . . . . 8
5. Instrumentation, Equipment, and Field Procedures. 14
6. Results and Discussion. . . . . . . 19
Temperature . . . . . . . 23
pH. . . . . . . 31
Conductivity. . . . . . . . . . . 39
Dissolved oxygen. . . . . . . . 43
Calibration . . . . . . . 50
Summary of data for all parameters. . . . . 51
References. . . . . . . . . . . . . . 56
v
-------
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
FIGURES
Page
Instrumentati on read out devi ces. . . . . . . . . . . . . . . . 16
Enclosure designed for Martek Probe. . . . . . . . . . . . .. 17,
Side view of enclosure for Martek Probe. . . . . . . . . . . . 17
Martek Probe after tests. . . . . . . . . . . . . . . . . . .. 19
Martek Probe and enclosure after tests. . . . . . . . . . . . . 20
Martek probe after tests. . . . . . . . . . . . . . . . . . . . 20
Hydrol ab Probe after tests. . .' . . . . . . . . . . . . ., . . . 21
Ocean Data Probe after tests. . . . . . . . . . . . . . . . .. 21
Hydrolab Probe after tests. . .' . . .,.. . . . . . . . . . '. . . 22
Ocean Data temperature vs. standard temperature,
for Data Set 111. . . . . . . . . . . . . . . . . . . . . . . 25
Ocean Data temperature vs. time, Data Set III . . . . . . . . . 26
Martek IV temperature vs. standard temperature,
Data Set III. 0 0 . . . 0 . . . . 0 . 0 0 . o. . . . . . . .. 27
Martek IV temperature vs. time, Data Set I I 10 . . . . . . . . . 28
Hydrolabtemperature vs. standard temperature, .
Data Set III. . . . . . . . . . . . . . . . . . . . . . . . . 29
Hydrol ab temperature vs. time, Data Set I II . . . . . . . .. . 30
Ocean Data pH vs. standard pH, Data Set I I I . . . . . . . . . . 33
Ocean Data pH vs. time, Data Set III. . . . . . . . . . . . . . 34
Martek IV pH vs. standard pH, Data Set III. . . . . . . . . . . 35
Martek IV pH vs. time, ,Data Set III . . . . . . . . . . . . .. 36
Hydrolab pH vs. standard pH, Data Set III. . . . . . . . . . . 37
Hydrolab pH vs. time, nata Set III. . . . . . . . , . . . . . . 38
Martek Conductivity vs. standard conductivity,
Data Set III. . . . . . . . . . . . . . . . . . . . . . . . . 41
Martek conductivity vs. time, Data Set III. . . . . . . . . . . 42
Hydrolab conductivity vs. standard conductivity,
Data Set III. . . . . . . . . . . . . . . . . . . . . . . . . 43
Hydrolab conductivity vs~ time, Data Set III. . . . . . . . . . 44
Ocean Data DO vs. standard DO, Data Set III . . . . . . . . . . 46
Ocean Data DO vs. time, Data Set III. . . . . . . . . . . . . . 47
Hydrolab DO vs. standard DO, Data Set III . . . . . . . . . . . 48
Hydrolab DO V$. time, Data Set III. . . . . . . . . . . . . . . 49
vi
-------
Number
1
2
3
4
5
6
TABLES
Page
Least Squares Regression of Observed Temperature Against
a Oefi ned Standard Temperature. . . . . . . . . . . . . . . .
Least Squares Regression Analysis of Observed pH Against
a Defined Standard pH . . . . . . . . . . . . . . . . .
24
. . . 32
Least Squares Regression Analysis of Observed Conductivity
Aganist a Defined Standard Conductivity. . . . ~ . . . . .. 40
Least Squares Regression Analysis of Observed Dissolved
Oxygen Against a Defined Standard Dissolved Oxygen. .
. . . . 45
Summary of Statistical Test for Null Hypothesis. . . .
. . . . 52
Data Set IV, Calibration Check of Probes After Cleaning. . ... 55
vi i -
-------
ABBREVIATIONS AND SYMBOLS
LSF
R2
-- least squares fit
-- coefficient of determination
CI
00
-- confidence interval
pH
-- dissolved oxygen
~- negative logarithm of the effective hydrogen-ion
concentration
MK
HL
-- Martek probe
-- Hydrolab p~obe
00.
NOP
-- Ocean Data probe
-- number of data points used in statistical analysis
T
C
~- observed temperature
-- observed conductivity
Tr
Cr
-- reference temperature
b
-- reference conductivity
-- slope of least squares fitted curve
ab
-- standard deviation of slope
viii
-------
SECTION 1
INTRODUCTION
This research project was implemented to provide input data to a longer
term effort aimed at developing a water quality monitoring system that can be
remotely deployed up to 45 days.
The first phase of the research was devoted to a literature survey and the
potential problems that may arise in deployable sensor design. Possible
methods to alleviate these problems were reviewed. It was found that no large
scale effort in the area of long-term monitoring with in situ sensors has been
mounted, due primarily to the many problems that are encountered when probe
packages are left unattended for extended periods of time.
The primary problems encountered are biological fouling and the lack of
long-term stability of the sensors, with chemical and physical degradation
contributing to a lesser extent.
Little can .be done within the context of this project to combat chemical
degradation (i .e., corrosion) since it would involve a redesign of the probes
to eliminate those surfaces that are prone to corrosion or the design of new
materials that are corrosion resistant. These are typical of tasks that would
be undertaken in a longer-term program.
Physical degradation (i.e.,collection of physical debris on the probe
surfaces and within the cylindrical housings) can be significantly reduced by
a redesign of the probe package to prevent foreign substances from coming into
direct contact with the active probe surfaces.
Biological fouling (i.e., growth of organisms on the active probe
surfaces) is the most serious problem and also the most difficult to combat.
The literature search resulted in consideration of several possible methods to
reduce biofouling: 1) in situ ultrasonic cleaning of the probes, 2) in situ
generation of sodium hypochlorite from sea water, 3) local application of
heat, 4) light-proofing the environment to suppress algae growth,S) genera-
tion of a local "kill zone" using toxic antifouling paints, 6) periodic
injection of a biocide, and 7) incorporation of toxic compounds into the
dissolved oxygen membrane.
Of these seven potential measures to eliminate b10fouling, the local "kill
zone" approach appears to be the most feasible and simplest to implement. It
requires a minimal increase in power requirements and has been proven
effective in a limited number of field tests.
1
-------
The test phase of the project was designed to determine the performance of
off-the-shelf water quality sensors in a realistic aquatic environment for an
extended period of time. Tests were accomplished without any adjustments to
the sensors after initial calibration and deployment except with respect to
the circulator on the dissolved oxygen sensor. In order to test the "kill .
zone" hypothesis, a container was designed to enclose one of the probe
packages, and a circulating system was incorporated in the "new" system to
periodically flush the container to bring a fresh water sample into contact
with the probes. The inside of the container was coated with an antifouling
paint that contained tributyltin fluoride as an active toxic agent. A
long-term (6 to 8-weeks) test of the probe packages in a realistic aquatic
environment was undertaken. These first tests were restricted to a fresh
water site; future tests will be conducted in saline and brackish environ-
ments. Four water quality sensor probes (a Hydrolab Surveyor Model 60, two
Martek Mark IV and an Ocean Data Model lOlA) were used in the field tests.
. Upon completion of the field tests and calibration checks, the long-term
probe performance was compared to simultaneous parameter measurements using
laboratory standards. The long-term performance of the sensors enclosed in
the toxin-coated container was compared to that of the other probes without
the "kill zone."
A major aim of this. project was to access the ability of these sensor
packages to maintain calibration in a realistic aquatic environment.
It was determined that both environmental degradation and the lack of
long-term stability of the sensors are major factors that must be overcome
before a successful depl oyabl e water quali ty monitori ng system can be
developed. .
Several tasks were implemented to identify and analyze those factors that
determine the performance of off-the-shelf water quality sensors deployed in a
.realistic aquatic environment for up to 6 weeks and to improve sensor qesigns.
These tasks included:
1.
a literature survey of problems and solutions associated
with biofouling of sensors to establish design modification to
existing sensor packages,
testing of a method to eliminate biofouling,
2.
3.
long-term (6- to 8-weeks) testing of the probe packages in a
realistic aquatic environment, and
comparison of long-term probe performance to simultaneous parameter
measurements using laboratory standards
4.
2
-------
SECTION 2
CONCLUSIONS
Laboratory and field tests show that uncertainties in the performance of
off-the-shelf water quality sensor. packages preclude their use for unattended
long-term field deployment. The following general conclusions about the
sensors were drawn:
1.
The Ocean Data sensor package is the most susceptible to
physical damage. Extreme care must be taken in handling
and in operations. The other sensor packages (Hydrolab
and Martek) are essentially the same and cause some
concern in calibration procedures.
For unattended, long-term deployment (greater than 10 days)
reliability, stability, and maintainability of the sensors
are unsatisfactory. The most significant problem is in
circulator failure in the Dissolved Oxygen probe~
2.
3.
There is a considerable stabil'izing period required for the
sensor packages to adjust to environmental conditions.
After a period of time, they appear to. stabil i ze andremai n
stable until major difficulties occur, e.g., circulator
failure.
4.
New and reliable calibration programs must be designed and
implemented for sensors used in long-term deployment.
(Sensor manufacturers do not provide adequate calibration
procedures. )
Biofouling in fresh water appears to be controllable.
5.
6.
Martek and Hydrolab temperature sensors perform reasonably
well; the Ocean Data temperature sensor performs well for
about 2 weeks and then begins to deteriorate.
The Hydrolab pH sensor is fairly stable, more so than the Martek
or Ocean Data, but all three sensors show some tendency to be
affected by time and/or fouling. The Martek seems to be affected by
time alone since biofouling was effectively eliminated. Hydrolab and
Ocean Data would probably perform better if kept free of biofouling.
However, elimination of the drift characteristics, which are probably
due to the inherent design of the circuits, would require design of a
more stable system.
7.
3
-------
1__-
8.
Conductivity measurements are very erratic. No definite
conclusions could be drawn except that measurements of
freshwater conductivities would be very unreliable due to the
l"arge uncertainties involved.
9. The Hydrolab and Ocean Data Dissolved Oxygen sensors perform
well as far as linearity and stability are concerned.
However, the large uncertainties involved would introduce an
error of 20 percent or more.
These general conclusions are substantiated by more detailed analysis of
each sensor system. The variables measured over the extended period are
discussed below.
TEMPERATURE
Both Martek and Hydrolab temperature sensors give good performance over a
41-day period of exposure to a freshwater environment at a temperature range
of 27 to 39°C.
The Ocean Data is stable for a 2-week period but tends to be sporadically
erratic for the next 2-week period. After thirty days of exposure, the Ocean
Data temperature sensor becomes very unreliable.
All three probes are readily calibrated for good linear tracking over the
temperature range.
pH
The Hydrolab pH sensor is stable with respect to time and fouling.
The Martek pH sensor is stable at the lower end point but at the upper
calibration point tends to drift as a function of time.
The Ocean Data pH sensor is fairly stable if kept clean, but is affected
by fouling, especially at the lower calibration point.
The pH sensors for all three probes will track within to.5 pH units of the
calibration curve existing at any particular time.
After a period of adjustment, the Hydrolab and Martek pH sensors remain
stable for 12 days in freshwater. The Hydrolab pH sensor may be more
sensitive to environmental degradation than the Martek, but no overall time
dependence was observed for either unit.
CONDUCTIVITY
Conductivity is difficult to measure accurately in freshwater because the
uncertainty in the observed values is often of the same order of magnitude as
the range of values being measured.
4
-------
At high conductivities, the uncertainty can be reduced to about 10
percent or less.
Due to the large uncertainties, it is difficult to calibrate the sensors
accurately.
Assuming that the conductivity of the water to be tested does not change
significantly with time, the average of the sensor readings over a period of
time closely approximates the actual conductivity. However, if for some
reason the conductivity were to change suddenly due to a sudden change in
local water conditions (e.g., the passing of a wet weather front), the sensor
would not be sensitive to the change unless the change were of the order of
magnitude of several standard deviations. Even then it would require some
time period after the ~hange before the change could be statistically
determined. .
The standard deviation of the conductivity difference is of the order of
50 micromho/cm which will create a large relative error for any given Hydrolab
measurement in water of low actual conductivity.
Tests in various environments, from fresh water to saline, will tend to
eliminate some of the difficulty in assesssing the conductivity sensors.
DISSOLVED OXYGEN
The Hydrolab and Ocean Data Dissolved Oxygen sensors track linearly over a
dissolved oxygen range of 0 to 20 and 0 to 17 mg/l respectively.
The indicated dissolved oxygen values will be within :1:2.0 mg/l of the
actual dissolved oxygen concentration.
The Martek Dissolved Oxygen sensor consistently deviates from the
standard. This may be due to the calibration procedure, which is a I-point
calibration. The linearity of the sensor response cannot be guaranteed.
BIOFOULING AND "KILL ZONE"
A "kill zone" is an easily implemented mechanism for reducing biofouling
by using a housing coated with an antifouling paint.
5
-------
SECTION 3
RECOMMENDATIONS
Many physical and electrical problems were encountered in calibrating the
sensors. The Ocean Data package, in particular, was very unwieldy and easy tc
damage. If the probes are to be deployed unattended for long periods of time
the initial calibration should be refined. Suggested refinements are:
1. Multiple point calibrations. Present procedures require
only one point for calibration. To ensure more accurate
data collection, several calibration points should be
used.
2. The calibration should not be a "one-time" calibration.
Several readings of the sensor to be calibrated, taken
.over a period of several hours, should be averaged and
the sensor set to the average value. The value of the
parameter being measured should be constant.
A continuation program should include the following procedures:
1. Develop and test calibration techniques fnr each sensor under
various environmental conditions.
Examine the feasibility' of increasing the number of parameters
that may be monitored.
3. Develop and evaluate new sensor designs to meet long-term
deployment requirements.
2.
4.
Deploy and evaluate the performance of water-quality sensors
under various field environments.
5. Under various field environmental conditions, evaluate individual
sensors and/or sensor packages available from other manufactur-
ers.
6; Continue a state-of-the-art review of sensor design technology,
including a review of the latest work being done in environ-
mental, medical, and space sciences.
7.
Initiate a study into DO membrane structures, especially in the
area of impregnated membranes using toxic, antifouling materials
as the impregnating agent.
6
-------
Develop a field testing laboratory, or mobile unit, to evaluate
all sensors to be considered for long-term deployment.
In studies of the "kill zone" concept for controllingbiofouling, the
design of a more reliable, algae-resistant probe and probe housing should
include the following characteristics:
1.
2.
8.
Requirements for design of circulating system for remote water
quality sensors
a.
low-current d.c. submersible pump
d.c. power source, probably solar assisted
b.
c.
non-corrosive structural material
d.
to reduce current drain, the power supply should have
.' low hydraulic resistance
e.
. complete change of water in minimum possible time
flow regime should be turbulent to obtain good mixing and
thorough flushing of "old" water sample
f.
the design of an optimum probe arrangement should be investigated
to accomplish characteristics d and e
resistance to biofouling or corrosion
g.
h. must not fail to operate for at 1 east a 6-week peri od in vari ous
water environments
i.
must circulate water about the end of the dissolved oxygen sensor
Antifouling compounds
a.
scientific study of leach rates and their effect on sensor
perfonnance
study of the different Qrgano-metallic compounds
b.
c.
detennination of the optimum thickness of paint coatings and
surface roughness
7
-------
SECTION 4
BACKGROUND
Early attempts to remotely monitor river water quality resulted from
increased demand for water due to population and industrial growth. Planners
learned that limited water resources could be better managed with the more
complete knowledge of river processes afforded by continuous monitoring.
Thus, in the late 1950.s the effort to develop remote river water quality
monitoring systems was greatly accelerated in both the United States and
Europe. Davies (1972) describes the European effort which attained many
operational 'goals in the late 1960.s. Typical of the European effort was work
by Cooke and Woodward (1970) who developed a monitoring system for the River
Trent. They measured temperature, dissolved oxygen, pH, and conductivity at
numerous stations located to provide data on the pollution dispersion
capabilities of the river. The problem of environmental degradation of the
sensors was alleviated by weekly sensor cleaning and semimonthly calibration.
Best (1974) reports similar objectives and results from research-done in
Scotland. .
One early American effort was the ORSANCO project documented by Klein et
a1. (1968). ORSANCO, operated on the Ohio River, was begun in 1960 and
consisted of 14 stations monitoring temperature, dissolved oxygen, pH,
conductivity, redox potential, and solar radiation. Like its European
counterparts, this research was designed to augment the management
alternatives of regional planners so that new industries might be located in
areas with the greatest potential for pollution dispersion. The problem of
environmental degradation was bypassed because of the stations. easy
accessibility for the frequent cleaning and maintenance of the sensors.
Palmer and Izatt (1970) reported the use of submersible recording water
quality sensors during work for the Ontario Water Resources Commission. A
one-quarter inch (6 .35-mi 11 imeter) magneti c pi pe was used to record water
quality data in a project to investigate chemical and physical characteristics
of a section of Lake Ontario, Canada. Problems of biofou1ing were alleviated
through frequent sensor cleaning.
West and Floyd (1976) designed and operated a system for gathering
baseline data on streams. Although the water quality parameters measured were
not different from previous systems, the field instruments were powered by
battery and solar sources. Previous inland systems had utilized power via
transmission line; however, power requirements for field operators can now be
met by existing technologies (LeBarge Inc., 1976, and Hydrolab Inc., 1975).
Physical and biological influences impaired the performance of some sensors
but weekly visits to cleaning and calibration provided for reliable results.
8
-------
Smith (1973) designed a buoy system for monitoring temperature, dissolved
oxygen, and conductivity in Narragansett Bay, Rhode Island. The data buoy was
operated for over 100 days during which time severe problems of biofouling
were encountered. Biofouling was caused primarily by Bryozoa and Hydroids,
which are animals, not vegetation. Growth rate of these organisms increased
as wat~r temperature increased. The marine growth limited permeation of
dissolved oxygen through the probe membrane causing inaccurate and unreliable
readings. The electrolyte solution within the membrane became contaminated
with suspended sediments, which might have affected probe performance. The
duration of reliable dissolved oxygen readings was not explicity determined.
Temperature thermistors and conductivity electrodes were covered by marine
growth, but their performances were not impaired.
. The U.S. Geological Survey, Baton Rouge, has devised a means of
eliminating biofouling and physical damage to sensors at two continuous
recording water quality stations in Louisiana (R. Lewis, personal
communication, 1978). A submersible pump lifts water into as-gallon
(19-liter) sump through which it continuously flows. Water quality sensors
are submerged in a sump located in a small field station. Water quality
readings are periodically made after a timer activates the sensors. After the
readings are determined, a second timer activates another pump that discharges
a hyper-chloride solution into the container. The resulting acidity kills any
organisms that might have attached to the sensors. Within minutes the
flushing action of the submersible pump replaces the acidic solution with
river water, thus restoring the desired water medium before the next sensor
observations. .
The literature on remote water quality monitoring reveals that most
stations are remote mainly in the sense that technicians are not manually
recording data. Most stations are.accessible to power via transmission line
and are easily accessible by car or truck. All stations recorded data on
strip charts, paper or magnetic tape, or via local telemeterinqdevices.
Of the environments in which monitoring systems have been tested,
Narragansett Bay (Smith, 1973) is most similar to conditions in the Louisiana
wetlands: experimental sites are not accessible by land or power lines, and
rapid biofouling is likely to occur in the estuarine environment. However,
Smith did not test the duration of reliable sensor performance.
In previous work conducted at LSU (Gasperecz, 1978) the difference between
Hydrolab and auxiliary dissolved oxygen data was of primary interest, because
of the general agreement among current researchers that the longevity of the
dissolved oxygen probe is the limiting factor affecting the duration of
reliable oxygen or sensor performance (Lewis, U.S. Geological Survey; Smith,
Louisiana State University; and West and Floyd, U.S. Army Engineer Waterways
Experimental Station, personal communications).
The importance of proper water circulation around the dissolved oxygen
probe was clearly illustrated in the data obtained in the LSU experiments. A
sharp change was noted in the difference between Hydrolab measurements and
9
-------
auxiliary measurements corresponding to the point at which the circulator
ceased functioning. The circulator failure was probably caused by a
combination of power drain and physical resistance due to sediments and
organic detritus in the area of the pin on which the circulator rotates.
Negative difference values indicated lower dissolved oxygen measurements for
the Hydrolab than the auxiliary sensor (due to oxygen depletion at the
Hydrolab probe membrane). It is clearly evident that water circulation is
essential for reliable dissolved oxygen measurement and that circulator
failure limits the longevity of accurate sensor performance.
The LSU data relating to performance of the dissolved oxygen subsystem
form a base against which future comparisons of probe performance can be
made.
It is noteworthy that the data presented by Gasperecz were obtained in mid
to late summer. This is significant because the growth rate of aquatic
organisms is highest during this period due to increased insolation and water
temperature. The estimates of duration of accuracy presented are based on
data derived during the period when the rate of biofouling would be at its
peak and are probably shorter than those that would be derived during periods
of cooler weather.
The major technical problems encountered in establishing a remote,
. long-term water quality monitoring system are biological fouling of the
sensors and stability of this sensor with respect to its calibration.
Chemical fouling (corrosion) and physical fouling (deposition of mud, sand,
debris, etc., on the sensors and sensor housing) are present but do not pose
as serious a problem as biological fouling. The main problem of physical
fouling is interference with the dissolved oxygen probe circulator. Early
failure of the circulator is usually due to excessive physical degradation.
The primary problem of biological fouling is the growth of aquatic
organisms on the active surfaces of the sensors. Fouling of non-active
surfaces presents a nuisance but does not interfere with parameter
measurement. Several surveys and laboratory tests were initiated to determine
possible methods to alleviate or eliminate biofouling of the probe surfaces.
Possible solutions were considered in LSU laboratory projects and are
summarized below.
1.
In situ ultrasonic cleaning of probes
A feasibility study of this method indicates that ultrasonic cleaning
could be successfully employed. Such a system has been used by Lewis of the
USGS, Water Resources Division, Baton Rouge, Louisiana. However, Mr. Lewis
had the advantage of direct a.c. line voltage as the power source, unlike the
situation of a remote sensor utilizing a d.c. power supply. One of the major
disadvantages of ultrasonic cleaning is the relatively large amounts of power
required compared to the power requirements of the sensor package itself. In
addition, the need to convert the low voltage d.c. source into a high
amplitude a.c. supply, the potential damage to the dissolved oxygen probe
membrane, and the additional mechanical complexity due to the addition of
10
-------
the ultrasonic transducer, make the use-of ultrasonics less attractive. A
laboratory test at LSU resulted in giving low priority to this mechanism for
probe cleaning.
2.
In situ qeneration of sodium hypochlorite from sea water
This method has been used successfully in industrial applications by
Connolly (1973) and Lamb (1972). The passage of sea water through an electro-
chemical cell generates sodium hypochlorite and hydrogen gas. Such a cell
could be used inside a closed container t~ generate a toxic level of
hypochlorite. Although this method would probably work well to prevent
biofou1ing, there are several disadvantages: the method will work only in sea
water and would not be applicable to fresh water monitoring, the hydrogen gas
generated as a byproduct would have to be removed, the resultant chemical
products may interfere with the probe readings, and a circulating system would
be necessary to periodically bring a new water sample into the container.
These considerations make in situ generation of sodium hypochlorite a feasible
but unattractive solution.
3. Local application of heat
It has been demonstrated in industrial applications by Anderson and
Richards (1966) that an increase of approximately 50°F above ambient
temperature for 30 minutes per week will prevent growth of mussels and
barnacles, although this may not be true for algae growth, the predominant
form of probe biofou1ing. Since most algae are active within a relatively
narrow temperature range, a temperature increase of 50°F would probably
increase the local temperature above the maximum tolerable temperature for
most algae. It is not known how often the heat should be applied or how long
each application should be in order to prevent algae growth. Potential
drawbacks to this method are excessive power requirements, if frequent heat
applications are necessary,. and potential damage to probes.
4. Liqht-proof environment to suppress photosynthesis
Since algae are plants, they require light to grow and reproduce.
Enclosing the probes in a light-proof enclosure should, therefore, be
effective in suppressing algae growth. This would require a circulation
system to periodically expose the probes to the water sample to be tested.
The disadvantages of this method are that a circulation system would be
needed, and that peri odi c f1 ushi I1g of the enclosure wou1 d expose the probes to
fresh algae and thus reduce the effectiveness.
5. Local "kill zone" produced by toxic antifouling paints
. -
Antifouling paints, in particular those based on organometallic tin
compounds, have been found effective in eliminating biofou1ing in
investigations by Caprari and Rascio (1972), Saroyan (1968), and Rohling et
a1. (1975). Since the probe surfaces cannot be directly coated with an
antifouling paint without interfering with probe performance, it is proposed
to generate a "kill zone" at the probe surface by enclosing the probes in a
container that has been painted on the inside surface with a toxic paint.
11
-------
There should be no interference with probe readings as the container would
be flushed before each reading to bring in fresh water. The leach rate of the
tin into water is slow enough so that. the measurements would be completed
before a high level of tin could accumulate. The leach rate would have to be
fast enough, however, so that a toxic level could be attained before the next
measurement is made. Since a faster leach rate reduces the effective fouling
resistant lifetime, it may be necessary to develop a paint formulation that
provides both the desired effective lifetime and sufficiently high leach rate.
The disadvantage of this method is that a circulation system would be
required, thereby increasing the complexity of the total system.
6. Periodic injection of biocide
The use of a general purpose biocide such as sodium hypochlorite has been
used successfully in industrial applications by Connolly (1973), Lamb (1972),
and Anderson and Richards (1966). A biocide should also be effective in
eliminating biofouling of the probe package. If the probes are enclosed in a
container into which the biocide is periodically injected, a sufficient
concentration will produce a toxic level. This procedure has the disadvan-
tages of requiring a circulating system for the water and a biocide injection
system. Also, the long-term effects of the biocide on probe operation would
have to be investigated.
7.
Incorporation of toxic compounds into dissolved oxygen membrane
The parameter most sensitive to biofouling is the dissolved oxygen probe
because of algae growth on the membrane surface which interferes with the
migration of the oxygen molecules across the membrane. The U.So Navy has
developed structural plastics that have a toxic compound incorporated directly
into the structural matrix, thereby making the material highly resistant to
bi ofoul i ng. These studi es. were reported by Ell er et a 1 0 (1970) and Dyckman et
ale (1974). A similar procedure should be applied to the dissolved oxygen'
membrane to render it resistant to biofouling.
Of the seven possible methods of reducing biofouling discussed above, the
use of antifouling paints to generate a local "kill zone" is the most feasible
because it is the simplest to implement, requires a minimal increase in power
requirements, and has been proved effective. In addition, the container used
to generate the "k ill zone II wi 11 a 1 so serve as a deterrent to phys i ca 1 foul i ng
and protect the probes from excessive wave action.
Antifouling paints have been available for many years, but their
performance has been questionable (Smith, 1973). The Uo So Navy has developed
coatings of optical quality (transparent) to inhibit growth of aquatic
organisms on submarine periscopes (Montemarano and Dyckman, 1972). These
organometallic polymers (aMP) have been tested for up to 5 months with
complete antifouling performance. Use of an OMP to coat the sensor's
non-probe components will substantially reduce overall plant and animal
growth.
12
-------
More recently, the U. S. Navy has developed antifouling composite
materials designed for utilization in construction of buoys, seawater piping,
shore structures, etc. (Montemarano and Cohen, 1976). These organometallic-
based composite materials have demonstrated complete antifouling performance
for up to 34 months. Utilization of one of these composite materials for the
construction of the structural components of a water quality sensor would
effectively eliminate biofouling on all non-probe surfaces.
The use of antifouling coatings and composite materials could solve the
biofouling problem on all parts of a water quality sensor with the exception
of the water quality probes. The probes' sensing components (dissolved oxygen
membrane and cell, conductivity electrodes and thermistors, temperature
thermistor tube), as they presently function, cannot be coated with or
constructed of antifouling materials and still maintain operation. A means of
preventing organism attachment to the probes must be developed to safeguard
water quality sensors completely from biological degradation.
Finally, a workshop sponsored by the U.S. Environmental Protection Agency
in Las Vegas, Nevada (Wruble, Koutsandreas and Pijonowski, 1978), substanti-
ated the need for new research and development efforts in the area of
deployable sensor development.
13
-------
SECTION 5
INSTRUMENTATION, EQUIPMENT, AND FIELD PROCEDURES
PROBE PACKAGES
Four complete probe packages. were obtained for use in this research: a
Hydrolab Surveyor Model 60, two Martek IV's and an Ocean Data Model lOlA.
Each of the packages is capable of measuring temperature, conductivity, pH and
dissolved oxygen. The Hydrolab is the property of Louisiana State University,
the two Martek's were obtained on loan from the U.S. Environmental Protection
Agency, Las Vegas, Nevada, and the Ocean Data Display unit was borrowed from
the Ocean Data Corporation for 4 months. .
However, because of equipment need of maintenance and repair, an initial
12-day test using only the Hydrolab and one Martek probe was conducted. Also,
before long-term testing could begin, one of the Martek probes developed
problems in the pH and dissolved oxygen circuits. Thus, the long-term test
utilized only three probe packages.
LABORATORY STANDARDS
The measured parameters were compared to standard values as measured by
the following methods:
Conductivity--The conductivity of the water in the proximity
of the probes was measured by a HACH portable conductivity
meter, Model 163000. The calibration of the portable
meter was checked periodically.
Rtl--The pH of the water in the proximity of the probes was
measured by a Scientific Glass Instruments, Inc., digital
mini pH meter. The meter was calibrated prior to each
measurement of pH.
Temperature--The temperature of the water in the proximity
of the probes was measured with a laboratory
grade mercury thennometer.
Dissolved oxyqen--Dissolved oxygen in the proximity of the
probes was measured by a Winkler test on water samples
taken before and after the sensor readings were recorded.
14
-------
The Winkler test was performed according to the procedure
described in "Standard Method for the Examination of
Water and Wastewater," (APHA 1976).
AQUATIC ENVIRONMENT
To simulate actual field conditions as closely as possible, yet still have
a convenient, easily accessible test site, the comparative tests were run in a
university fresh-water research pond operated by the LSU Department of
Agriculture. The pond is located in a university research area that provides
protection from vandalism. The pond is roughly 900 feet (275 meters) long by
200 feet (61 meters) wide and has a maximum depth of approximately 5 feet (1.5
meters). The pond is sufficiently rich in algae and other aquatic organisms
to provide a realistic sensor environment. .
MOUNTING PLATFORM
The sensor packages were mounted in a physically stable configuration that
also provided for easy access for data readings and any required maintenance.
For this purpose, a platform was constructed and erected in the research pond.
The readout consoles and power supplies were placed on the platform
. shel ves (Fi gure 1) and the sensor packa~es were suspended so that the top of
the sensor housin~s were just below the water surface. Before data were
taken, the sensors were raised so that the active probe surfaces were just
below the water surface.
POWER SUPPLIES
Power supplies for the Martek probe sensors were supplied by two 12-volt
automobile batteries. . The Hydro1ab systems used nickel-cadmium batteries for
the sensors and a separate 6-vo1t battery for the dissolved oxygen circulator.
CONTAINER AND CIRCULATING SYSTEM
One of the Martek sensor pacakages was modified to test the proposed IIkil1
zone" method of reducing biological fouling. A container was constructed of
sheet aluminum to fit over the Martek IV sensor housing (Unit Number 2). The
container is cylindrical with a funnel shaped exit plenum to which is attached
a submersible 12 volt d.c. pump. The top of the cylindrical container has a
one-half inch (12.7 millimeter) gap to allow water to enter. The inlet is
covered with fiberglass mesh to prevent debris and large aquatic life, which
could damage the pump impeller, from entering the container. ~igures 2 and 3
show the container system.
The interior of the container is painted with two coats of Sears
Antifouling Bottom Paint with tributyltin fiuoride as the active toxic agent.
15
-------
Figure 1.
Instrumentation read out devices.
An automatic timer is used to activate the pump for 2 minutes every 4
hours in order to bring a new water sample into the chamber.
FIELD PROCEDURES
The following format was followed in acquiring all data points.
1.
Apply power to sensor packages and the Hydrolab dissolved oxygen
circulator. (Since the power supplies are depletable batteries,
energy is conserved by turning 0" the sensors only when data
are recorded.)
Raise sensor housings so that probes are just below water level.
2.
3.
Check that all circulators are operating.
Allow probe readi ngs to stabilize (approximately 10 mi nutes), to
alleviate the problems associated with warm-up periods.
4.
16
-------
--- ~i; -:1
;. 'it&~J
'. ",
..; ~
.....- ~
.' . ~~. ;
'~-.
- .
-.
...
Figure 2.
Enclosure des i gned for r4artek Probe.
The inside surface of this cone-shaped housing was coated with two coats of an
antifouling paint (tributyltin flouride) to create a "kill zone" for combating
bi ofouli ng of the sensors. The housi ng also contai ns a small pump to prov; de
periodic flushing of the enclosure.
':.~-" '.
Figure 3.
Side view of enclosure for Martek
Probe.
17
-------
5.
6.
Check pH and conductivity meters while sensors are stabilizing.
Take a water sample at the water surface between the probes, and
immediately fix the dissolved oxygen content according to the
procedures for the Winkler method as described in APHA (1976).
Use "Alum Flocculation Modification" if required.
7. Record Hydrolab readings for temperature, pH, conductivity and
dissolved oxygen.
8. Record values of temperature, pH and conductivity as measured by the
other instrumentation.
9. Manually operate circulatory pump for 2 minutes. With pump off,
record Martek IV readings for temperature, pH, conductivity and
dissolved oxygen.
10. Record Ocean Data readings for temperature, pH, conductivity and
dissolved oxygen.
11. Take a second water sample at the water surface between the probes
and fix the oxygen content as before.
12. Check all probe readings to ensure that readings are stabilized.
13. Turn off all power supplies.
14. Lower sensor housings to original positions.
15. Check to ensure that everything is firmly in place.
16. Deliver water samples to the laboratory for i~mediate Winkler
analysis and record the results.
Record data three times daily at intervals of approximately 4 hours (early
morning, afternoon and evenings).
18
-------
SECTION 6
RESULTS AND DISCUSSION
The data presented 1n this section were obtained from three sources: a
12-day test performed on the Hydrolab in three different water environments
(Gasperecz, 1978) (Data Set I), the 12-day initial test of the Hydrolab and
one of the Martek probes in a freshwater site (Data Set II), and the long-term
(6-week) test of the Hydrolab, f-1artek and Ocean Data sensors in a freshwater
site (Data Set III), the latter two testing periods being conducted in this
current project.
Before being calibrated and deployed, the Hydrolab and Martek sensors were
cleaned to ensure that no residual algae or sediment were on the active probe
surfaces. The Ocean Data required no cleaning.
After removal from the water after a 6-week period, the probes were as
shown in Figures 4 through 9. The Hydrblab and Ocean Data were significantly
Jt .
.r#'
..
..
Figure 4.
This probe was protected over the 6-week period by the housing shown in
Figures 2 and 3. Note the absence of biological growth on the sensor area.
19
-------
,.#!~ ".1 ~, ..;c .. ."-..... '. ~ 1°f,' ','." ',A, to "......, .. .'",.. ..u~....'»" ~. o. .. ".,... -..,. ...'
. "!: ',':', . \~".""j""_\."";.""'# 7;*\#..."\" ..4"1..':' " .,.,f-!~.' J.,;;~ \":~.'.:'~""'~::- -."... ."..r. ..
; - '.' ,~)" .... e" . #. ~.I._". )':~..,.',":." Iff" .,'-.'~~'iI.7'.~.:.:ft~!.... "';', ; '" ~~ . ~...~~i~.:~..;/I.~~.: :~-"-4 '.: l-~~
. '. " ".'''' '.."... .:... 't~... .f.~.:'..~/,.r;v."'}'.:~..~.. '.~ .~.'.'-. .",_~,r.; . .'o'!'>"~ "~.'. .".'.;.;""" '~.~.'.:."" .~.. '
.' . .,; ... ~ ".; ..~-'-. . .~:'il~~_;;'!~':';,:'" ~':!'~'J.:.... i"'~'" ~':.'." .... ~.~.. - ~,,'c,..:....~,~'.'~\ . ',;.:', .":l.. .,
;., '. . .~, ..,. ".' ....' .~# ..;~,.:,,~,;:1.'.r~',,~~,\ .,.:,!,,:.:.j'~ ..,..."7'~~"i...::..:.",~...........-; ~:. ,;.j1
~ , . -. ... ... ....~.. "...''f''1~ '., '}:..:';!;~:'j~~ t':'~"f:~:'. ,II ~"".,:.,.P !.~ .,...~t r '.";'''~ ~
. . 4: . ".~. '''''.''",,~..,.,.,., ~,'~.;.3~,~~" ~rt,~,;f'i;'-',~ '~~.....:~",:'f"1......~..'>~-..\:':~'r'...;~;........i:..'t
. ". . .." 0." f_,.,.-,.,,' . '. ". ".' '1/".. . ;..;.).' ,: ,I~.. .~. ~"..' ..~' '. . .';."-. '~«1 1'.tU.'.i.t.!!. i\t~1 fr!,,...' . '.';.'r ~,...~ ~......~~ . "~.-... ~,.,'~ ..:.' '~: . ;.:- , 4,' I
-,.' . ~;':I.~~';'-, ",,,,,!,,...t. .', ,\.' ,'-' 'tl!.'-t;"';:i.':;p.-:Y"~'-'...;" +;.u-.. '<..,:,1..' '.- ,,:,;, ..
'. "..- ';"C~p. "'; '~':.'~'- . ~:r;~~.. ~,.~.,~. ."., "~~'"!:;if; ,," ..t",,~;, .' . {iiI
~;;hr;" ~:":;;~~~f';:.~"ft:;t~:i~'~(;~ ~"~:.' '.' ~"'" "~:.' ...,~ ", qi~~~~f~':.f~;;;,.~.>:.;: L,;i::;~;;~, ... .;~?=, 'Ii~
:~~". . '. ..''''. :i.,'~~~.~' V' ~~~,.,j -":[-:~ :....;;'.~':'.'t!'J...' ~'~, '~,.':..~:~. . .'~ .~''''.~~~~.';..'.'..';;'.'.~:''.'~,'< \;.'!>.,r<..:;:;:'i;;~'. ::~ '.. ':'ii-.; -', ,,::'" ;~. fJ. "~.'"
, '''''''''f.';o...'i'.;~~-''':<. .~('~,. .,,~... '''',,,.. ~::.:"""'"'.'l\'.....r'-''' ...,t", "'. .. ...' ,".-"., .."
"~~. .. ...';":; cJ.'>-~~~ {~"""M;~~"'~,::-': :';~ ~.:~>~.~. ~'~~P'~""- "..~~~~~;;;-4~~t;.;:"h"~;::U}~-;'ft'~.:'~~-:i' :~:~.~. ~"~:~'~':. \.-: ~I:~ -)~\,..
..... ~~.~, .. '. :;'~~" .' '~1r~.".," \,is;~~:,4;;£:,;;:.";,;:~i~.,..,:ii1 '~~t;~;!'~A:;~:, 8i..~':f
. ..~ ' . -. ..'" .- .~. 1:" .'~.JII:".. -..." ",,,,'.'...:;t'. ,''Co<, ."J,ht. ~';,1 p. ';' ..~ ~"',.I! ~.,.':'
"-, ..... ":. ~",... ,) ".'''~ ::I!':~.,~;;1~~:;""C;"::'._,~- ." -.,~;;..!.t-'''n.
. .. -, ..r ......~. f/I.' ~~\. .' .tt... '.-
-------
t$";~'4~
~~. ~'1'!':
>~~.,
.'. W-
o' ,
f!)f~ .~~
I'~$:." ,~~,'
.1..:..:" .
".~',
. ,~f$,. )
-III "
1~.:E~:
Figure 7.
Hydrolab Probe after tests.
."
I .?"
l.:t."
7
Considerable biofoul1ng is evident on this housing and probe system as
in the different gray scales on the photograph.
noted
21
-------
~, ~~I4'';;~t fi"'/,II,:'~",. ~, : ',);""':"",' :H,<", :/",'u~>:')'''''''', '~",-"', ", '-"'~,-,'\.,"'," "~'" " '~"'-'''''''.l
~ ,~~z"'-' '*'!' f~J...,~",if';';'a::::"~ ;~'''':.:U!'~;:~-~;:',~;.~,,~ '~'(~"""1'"..._..,~ i"~"'~i ;..).~~~~,:.::,'. ,;~. ,,:":'..,":.
~'e.~ i-i;I'~ J.":':ri::.~'J.,~~;: . ,...,-t:~''''~~~:~~If::J~~~',.~:":'''';;:'.,f4';:~:.r~;}
;;~~!~:.~-~ '&),' ~', ~~~~~I;J;~,',!~;£~ '~;~~.F ~~-~~,:--,,~~~l,~,'~~.~~,"~,2.,t~...if~, ':,'~-f~~,~~J]~~:,'~'~' \
~, :,h..,);.. ,,,,~, JJ;J." . '''-..1;0,.1..,.;, . :, ,"'.",,::..;01' iI"!!''''' !';;;"""~',(,~-".",~f':.'~'~'" ..,,:t;:';,;;,'. ~,~.,f ';;'~(:~,"/J;.:.i"~""':!"
':....1'';. -" ~ '" ~. t' :".~;':;";..." -",:,=~/~,*1II::<"':'''Y~''''':{'~""~,,, .:. .:-r,:, '. ;:':"! ._...' ,
:;;,~. /~~; ~g1::~'~~~'i;~~i~;;h~' ",:' '~~,~~~,,:-:;:~~~{ot.. ~1Ji:4;tr;j','!,I'J, " ;.;t..,~~..,
I'~'~~'" f:~;.,~":~;,'~~"';";;'~--;;' ",t'i::;~'~:",rit:i:~. '~, ;,"'~,~,~~~,", ,'~':,~~' ~~~~'~~1~':",' .;'
'. '~''!.~'~'4 :~' ~-m..r1- J*,,", ""', 'i;iff.;"'" " ,'!f!!~;". .., "~,,~'j,,.;........,.;- ~.~ .,
, ~.~,.",~.. ~.':'~, ..\..:~ "~ '!'~'''' ".... .:~*"'. .. ';"'i,:".~-w.:!.o.l~.t~~""'t, \. '. t,"':\.~ ". ..... ~~.;;. , 1," I'
-, -,!.;""," wl_"i" ','V' ''':..~'''_''~;''. >":1' ~",;:,.,.......,J'" ~'" ,.,~, ','" . ;' ...
-..,~.. ",......"..~ .,... .~ tt,' -.k.., ~'J\--'~~'r"""",,"\~:'. ..... ..
-:..4, .:,,:'" '.' ~.;*,-'..:....,....1~~-- " .... ~~"J."'~''rl1l.!ifl':' '~~.?I-~ ..,' ~ '.
~~:~:~.:.J'7....:~.t 7',~'!:'1.1~':.';1.v ~ ~.' .~~.. ~'\1.~~~~. .. u;.. A~~~. :~.,~~ "~',..~. .
>,-" """",,,,,,,,,,,,,,,,;,,,\.~~.., . ". ""if '~" ,!1.<...." 1. ';' "~~'"
~ ..; , -,.-...eo,.. ..... r' ;~!),t-"- t..' ," ...;..~--'\ .. 4. ...1 ~
,.. " ..' ': - -~ ... \.-..,... ~", .. ;','''''' , It ". ~~.'~ "',' , ,'-.' .
. . ..~.'" ':W-'''~", .'.'!" , '-.. ,4... ."' '. . .' .
.tt;;..., . ,.", ", . . ~ ".' :-"j,",~ \ .' ::~ .. ," ," ~,:::~~:; "" " ." ....." ~ . ~ Ii..:".. ,"-
:~~,'~ ,~...:, I .. ,,' ~.. '. .;..~..:' !. 'f.ij,' " . ,}t:~~'\\' .' . ,... I ,
~.. ~ IIIIIt ....'.. ".;1 1.,~ '.., ""',.' At'" ,,-," . ~ ..'
.of ' ,'.', ','; -.':..' '~'1, ,'I':'~'" " '.. 0( ';oi;' .
",;.....' '. ..' ~ ,':' " .'.~:.~.,ft. ,f:.," JAt .. "', "'~"..I,..." '
.~ ,':4 '," . , '. ,~~. ... ' ,,~"i .~f'... ..,' :J:,!, , '.. J1,:~, ...,. .
'!fJrl ' ' ," . t' -.' :.' ,tit,'r'~~:~~,!5i:',(,."':,>", '..oJ>, ~',""~~., '.
~,...8 ." ,..," ~ ',~J, ..:ti,' ,',J'.. ,~..,." '"'@,.. ... .~\. :'
I'.~"':t. ", ':,'.: ,'~, ,':J.~". ";"'..~,, ',.,'~.~.,-..J.:. .
',-V-.,' - '\...-1.;' '~'''''~''''~'''''''''' ',',J~""~'.
't.~;,", .'. .'.,.'. .~.. 'if?.: ~1';"~1~;'~4~"~~~<~". ~.:"I"~:N- .., .
..-:~.. - " '..,' ""'., 1U ., .,..' .."-.. '~'" .. .
......4I(j~". .-. .-;,. <:':~,~- \ ,.,;1:t.','ll' .;,.tJ'f'..-':.I\-
~-.--..-,~,... . .. ...., '. " ...~ ~.", ''f':'~'' .'
'~':~~~, I .'.~~\ ~.1:' ",:"t7J ..«: ~I."'.~."~~;'~~'~'~'~~:\~ ~:'-..~~...~.., -: .
~.~~ .~ ~ ~."'.t:~~~~~~,.,~,.,~\.~;" .,~~...
. ,.. \' , '; .' :Jo:~"/~""'" "",;~:;".<',",,~... .. .).-an...;
~i.~,.'ti....!J~ i I~' ~ .:::~,,.,,.'i,'L.;:~'~...-;.~~~/~:~~ ,..,-
... .' ...., ,'" r' ~,- -'~"-" .' ..,tOr.. . !!:-....,. "'IIoj,
':'"~-'~'~.'," { ;B';':"~ ',~ '''''''''''~''''.'"-':'~~:~...''':='''''\''':,,\,
~"".l';;,!,',~ t. " ~'p~~~..~~~l't~~~~~,~~~j~~~~f~"\
~~,~-';:~~j ! !\.. ~~~.t_~~~~~~~:~~~~~;~:~~~~~,I~.~.
'.
.
. -'
-
a
;
Fi gure 9.
Hydrolab Probe after tests.
Another view of the biofouling which occurred during the 6-week test period.
covered with algae growth and other aquatic organisms. The Martek, which had
been enclosed in the protective enclosure, showed no significant difference
from its initial condition except at the upper end of the housing which was
not completely protected by the antifouling paint. It is evident that the
antifouling paint is very effective in eliminating biofouling.. It 'is also
possible that the agitating action of the circulating water (2 minutes every 4
hours) contributed to the reduction in biofouling. Additional support for
this hypothesis is obtained from the observation of the Hydrolab sensors after
removal from the water.
Examination revealed that the surface of the Hydrolab dissolved oxygen
probe directly above the circulator was very clean and free of algae.. In
addition, the glass electJ"Ode of the pH probe was entirely covered with algae
where exposed to the water, except for a small spot (0.5 cm x 0.5 em) on the
side facing the center of the probe package where the circulator impeller is
located. Relatively clean areas on all the probes were observed at those
locations' where the probe surfaces were enclosed by a plastic support housing
through which the sensors were mounted. The dissolved oxygen membrane was
covered with a thin film of foreign matter which easily wiped off.. The
membrane was still transparent but had a cloudy appearance.
The Ocean Data pH, conductivity and temperature sensors were all
completely covered by algae and other foreign material. The circulator,
instead of a circular impeller as in the Hydrolab, consists of a vibrating
22
-------
wedge positioned about one-sixteenth inch (1.59 millimeters) from the
dissolved oxygen membrane. The dissolved oxygen sensor is positioned away
from the other sensors and attached to the outer housing of the package, and
the effects of the vibrator would not be felt by the other probes. The
dissolved oxygen membrane was covered with a thin film of foreign material
similar to the Hydrolab. .
In the following discussion, the performance of each parameter is
discussed separately.
TEMPERATURE
Of all the parameters measured, the temperature was the most stable. The
results for Data Sets I, II and III are presented in Table 1 and plots for
Data Set III are given in Figures 10 through 15.
The slopes of Least Squares Fit (LSF) ranged from 0.90 to 1.065 for all
data.* .
Coefficients of Determination, R~J were greater than 0.93 in all but two
cases.
The slopes of the curves (Figures 10, 12, and 14) are not statistically
different from 1.0 at a 5 percent probability level except for two cases.
The 95 percent Confidence Interv~ls (CI) of the slope ranged from to.5°C
to t1.5°C. Seven of nine data sets had 95 percent Confidence Intervals
between to.5 and t1.0°C. '.
The Ocean Data temperature sensor performed very well for 15 days after
which time the sensor began to give excessively large deviations from the
standard temperature. This occurred twice for periods of about 3 days each,
separated by a period of about 7 days of normal operation. After 30 days, the
Ocean Data began to give increasingly high readings. This trend was
continuing as the test ended.
The Martek was removed after 15 days of exposure to check the dissolved
oxygen sensor. It was replaced after 1 day had elapsed. The temperature
sensor was not adjusted, but the apparent slopes of the temperature curves
changed from 0.955 to 0.902.
* In the discussion of other parameters, the abbreviations LSF, R2, and CI
wi 11 be used.
23
-------
TABLE 1.
LEAST SQUARES REGRESSION OF OBSERVED TEMPERATURE
AGAINST A DEFINED STANDARD TEMPERATURE
Summary of Least Squares Fit (LSF)
Data
Source
Range of
Temperature (OC) Slope"
Intercept
R2* .
0.95 CIt
NDP**
I, HL
freshwater 29-33 0.934 2.293 0.79 t1.0 69
I, HL
brackish 29-34 0.991 1.276 0.62 t1.5 78
I, HL
seawater 25-32 1 .065 -1.184 0.97 to.7 62
II, HL
freshwater 25-32 0.995 -0.728 0.93 t1.1 36
II, MK
freshwater 25-32 1.013 -0.638 0.98 to.6 29
I II, HL
freshwater 27-39 1.017 -1.300 0.97 t1.0 93
III, MK1*
freshwater 27-39 0.955 0.732 0.97 to.S 42
II I, MK2*
freshwater 27-39 0.902 1.930 0.98 to.9 48
I II, 00 ~
freshwater 27-39 0.986 -0.191 0.96 :to.7 41
* Coefficierit of Determination, R~
**Number of data points.
t Confidence Interval of the Slope
* The Martek p~obe was removed 1 day after 14 days exposure for a check of .
the dissolved oxygen sensor~ The MK1 data are for the first 14 days
exposure, and the MK2 data are for the 26 days exposure after the probe was
redeployed. The temperature sensor was not adjusted.
g After 14.7 days exposure, the Ocean Data temperature began to indicate
excess; ve devi ati on from the standard temperature. The data given here are
only for the first 14.7 days.
24
-------
a
a
.
a
::r
I I
Too = -.191 + .986Tr
+
.
,
+
+ +
+ +
+ +
+ **
...,.. :or ++
4 t
*:j:$/r +
:j: +.:f + +
++
+ + +++
+ .
~ . :t-:j:f: +
T
~+
T +-+
+ +
+
+ + +
. ++
+ +
+ +
a
a
.
"
"--"(Y')
U
t.:)
W
DO
"'--'a
.
::r
CL(Y')
:E:
W
~
a
a:~
~-
a:M
D
z:
a:o
WO
Uoo
ON
a
a
.
l(')
"25.00
28.00 31.00
STRNDRRD TEMP
34.00 37.00
(DEG C)
40.00
Figure 10.
Ocean Data temperature vs. standard temperature,
for Data Set III.
25
-------
a
a
.
Ul
U
C)
lLiO
O~
- .111
CJ
L1:
IT
00
20
0:"":
I-
(f)
N
0'1
10
o
(L
~-
,
li.J
I-
0:0
1-0
0:
07
--
+
--
+ +
++ +
-I-
+ +
't t
-t+
-j...
~ ++ + + +T + .1- T
-t+t( + iJ.-t++ + +
-I- + 1+ + * -&- .1- + .,1- +
-tt-+ + -&- .J.. .J.. +
. t + +
+ +
-I- .... -I-
+
i.
+
- -1-.
--:r--
-I- +
-I- -I-
.'--- --.-.---..-. -----
- - - - -
2
IT
W
UO
o
E) .
Ul
'0. UO
5.00
10.00
15.00
T I ME (D,S!
figure 11.
20.00 25.0U 30.00
FROM START OF TEST
35.00
Ocean Data temperature vs. time. Data Set III.
110.00
11:..00
-------
.
I'-
-('r)
U
C)
W
DO
.......0
.
~
o...('r)
L
W
I-
o
>C::
~-
~
W
I-
a:o
CIo
La:)
C\J
o
o
.
o
~
o
o
TMK = 1.93 + .902Tr .
/!
/
/
/+
+/
;7
('r)
o
o
.
U'1
~S. 00
28.00 31.00 34.00 37.00
STRNDRRD TEMP (OEG C)
40.00
-Figure 12. Martek IV temperature vs. standard temperature,
Data Set III. (The dashed lines are the 95 percent
confidence interval for the prediction of mean
temperature values for the Martek IV.)
27
-------
o
o
.
N
._._~
.-- ------
...
.
+
4- + +
+ jt.+ + + +
+ + it- +1
+-t+ + + -II- + + + +
... ... + ~ + +
+ + + + -I- + +-1/- +
.t I II I I --
it- + + .i- -t+
+ ... + ... + +
*.. +
... + +
... +
-I- -I-
+
-I-
- --- -. ----.-
-I-
+
1-
-. ... .-
- - .
5.00
10.00
15.00
TIME (0)'5)
20.00 25.IJO
FACJt1 5TI~RT OF
30.00
1EST
35.UO
110.00
I! :, . II Co
u
C)o
Wo
o .
~.-.
N
~
o
0::
0:0
00
z.
ero
I--
({)
10
o
(L..--.
:£1
W
I--
>0
.........0
N
~I
W
I-
a:
ero
:£0
CD
'0.00
Figure 13.
Martek IV temperature vs. time, Data Set IiI.
-------
o
o
.
o
::2"
o THL = - 1 .3 + 1.0 17Tr
0
.
-r"
u(T")
Q
W
00
,--,0
.
=r
o...(T")
~
w
r-
0
coC:
a:-
....J(T") /
0
a: A+
0
>-0 / /
:cO
.
ro
C\J /
/
0 /
0
.
U1
~5. 00 28.00 31.00 34.00 37.00 40.00
-STRNDRRD TEMP (DEG C)
Figure l~. Hydrolab temperature vs. standard temperature,
Data Set III. {The dashed lines are the 95 percent
confidence interval for the prediction of mean
temperature values for the Hydrolab.}
29
-------
o
a
.
ru
U
<...:)0
Wo
o .
~.--o
o
0:0
ITa
o .
ZD
0::
W I-
0 UI
0
Ie:
I (LI
:2:
W
1--
a
(Do
ITru
-I I
0
0:'
0
>-0
IO
.
r----.-
+
----
+
+
+
+ .
+ + t- +
+ + +
+ -t+ + + + + + * +
+ 'It- -II- + .t- 11- +
+ + + + -/- + +* + +
+ + + + + + -- -t-
+ -t+ -II- +
- I-t--I- - I I --t 1':
+ + +
+
+
+
+
- -------
+
---- -- -
- -
('")
10.00
5.00
10.00
35.00
'to.OO
l15.00
I~.OO
TIME (OI'S)
20.00 25.00 30.00
FROM START OF TEST
Figure 15.
Hydrolab temperature V5. time, Data Set III.
-------
pH
pH sensor analysis is presented in Table 2 for Data Sets II and III and
plots for Data Set III are given in Figure 16 through 21. The following ob-
servations are noted.
1. For all data sets, the slopes ranged from 0.464 to 0.937.
2. Coefficients of Determination ranged from 0.61 to 0.96.
3. All slopes were statistically different from 1.0 except in two instanceso
4. The 95 percent Confidence Intervals of the slopes ranged from t002 to
tOoS pH units. .
5. All three sensors exhibited an apparent change in slope after the passage
of Hurricane Bob. The Martek not only changed its slope but also the
level. The Hydro1ab had a relatively small change in slope while the
Ocean Oata decreased from 0.793 to 0.525. One hypothesis for this
apparent anomaly is as follows. Prior to Hurricane Bob, the pH values
were clustered about a 9.5 to 11.0 range, with a relatively high scatter
so that a linear trend is not readily apparent. It is possible that a
change in the calibration could have occurred during this period that
would go unnoticed due to the small pH range. After the passage of
Hurricane Bob, the pH values decreased for a period of time to as low as
705 so that a much larger range of data was available. The actual slope
at this time would then be more readily apparent. The pH sensors were
calibrated over a range of 7.0 to 10.0, the range expected to be
encountered in the lake. Assuming that the LSF of the data before
Hurricane Bob is an adequate representation of the actual calibration
characteristics, the following fitted values of pH are obtained:
Actual pH
7 10
Hydrolab 700 9.4
Martek (MKl) 7.5 907
Martek (MK2) 707 1001
Ocean Data 7.3 9.7
These values seem to be consistent with the initial calibration range of 7
through 10. It thus appears that the pH sensors all underwent a deterioration
in the calibration over the 41-day period. The only other explanation is that
the pH sensors are in some way affected by hurricanes. After the hurricane,
if it is assumed that the LSF of the post-hurricane data adequately represents
the sensor calibration at that time, the following fitted values of pH are
31
-------
TABLE 2. LEAST SQUARES REGRESSION ANALYSIS OF OBSERVED
pH AGAINST A DEFINED STANDARD pH
Summary of Least Squares Fi t (LSF)
Data Range R2*
. Source of pH Slope Intercept 0.95 CIt NDP:!:
II, HL
freshwater 9.0-10.0 0.503 4.462 0.83 :to.2 34
II, MK
freshwater 9.0-1000 00666 3 .496 0079 . :to 02 34
II I, H~1 9
freshwater 9.3-11 .0 00807 1.310 0072 :to.4 54
III, MK1#
freshwater 9.5-11.0 0.732 20364 0072 :to.3 35
I II, MK2 #
freshwater 9.3-10.5. 0.781 2.239 0.74 :to.5 12
I I I, MK3 #
freshwater 9.6-10.5 0.657 2.692 0.97 :to.3 4
I I I, MK4 #
freshwater 7.5-11.0 0.464 3.839 0.89 :to.4 36
I II, 001**
freshwater 9.5-11.0 0.793 1.767 0061 :to.S 49
II I, 002**
freshwater 7.5-11.0 0.525 4.544 0.86 :to.~ 38
I II, HL2 9
freshwater 7.5-11.0 0.937 -0.053 0.98 :to.S 38
~ Coefficient of Determination, R2
t 95 percent Confidence Interval of the slope
* Number of Data Points
9 The Hydrolab also appeared to change its slope after Hurricane Bob. HL1 is
for data prior to Hurricane Bob and HL2 is for data after Hurricane Bobo
# The Martek pH parameter exhibited an apparent correlation to the occurrence
of rainfall. The MK1 data is prior to the removal of the Martek package
after the first 14 days. The MK2 data is for the period from redeployment
at 1S days until the occurence of a light rainfall after a very dry June.
The MK3 data is for the period after the rainfall untn the onset of
Hurricane Bob which was accompanied by heavy rainfall. The MK4 data is for
the period after the passage of Hurricane Bob until the end of the test.
The pH sensor was recalibrated prior to redeployment.
** The Ocean Data also appeared to change its slope after Hurricane Bob. The
001 data ,i s for data pri or to Hurri cane Bob and the 002 is for data after
Hurricane Bob.
32
-------
IO
CL~
°
a:-
~
a:
o
zg
a: .
wO')
U
o
o
o
.
C\J
.....
o
o
pHOD = 4.544 + .525pHr
.....
.....
o
o
.
co
o
o
.
r-7.00
/'
./
./
++ "",
+ ,,/
"", .
./
./
11 . 00
12.00
+/ .
/
8.00
9.00 10.00
STRNDRRD PH.
Figure 16. Ocean Data pH V5. standard pH, Data Set III. (The
dashed lines are the 95 percent confidence interval for the
prediction of mean pH values for the Ocean Data.)
33
-------
OJ""'
I
IT
--'
o
0:0
00
>- .
ICD
10.00
--
-- 1--- --- ..--- --..--- --
+
+-\+
it- +
+ + \+ + +t +++
+ -It + + + + + -t '-
+ -14- -4 -lit- - -tit + + ~ if. 1-+
+ + ,+ + fr * + +
+ + + + + + ++
-II- + +
- .
.
-- ----- --
- ----- --
- -
o
a
00
0:
w 0:
~ 0
20
0:0
1--'-:
Ul,
Io
(Lo
5.00
10.00
15.00
T(ME tOYS]
20.00 25.00 30.uO
FROM START OF TEST
35.00
110.00
l15.00
Fi gure .17.
Ocean Data pH Y5. time, Data Set III.
-------
IO
(L~
o
>-
.......
~
Wo
1-0
a::: .
a: (j') .
L
o
o
.
.N
-
o
o
pHMK = 3.839 + O.464pHr
-
-
",'"
",-'"
o
o
.
co
o
o
.
r-7.00
8.00
9.00 10.00
STRNDRRD PH
11.00
12.00
Figure 18. Martek IV pH vs. standard pH, Data Set III. (The dashed
lines are the 95 percent confidence interval for the prediction of
mean pH values for the Martek IV. Values above 9 pH units on the
y axis were not used in determining the regression equation.)
35
-------
W
0\
I.
a
a
.
ru
I
---. -- ----- --- . .-..
+ +
. + + +
+ 'T + + T ... T
-*. +t + + + ++ +
f .~. + -fi' Ir"#- + +-t. ... ++ +
+ + + +
+ t -q.-P-
.t
+ ...
.
-t. + +
+ +
+ ... .+
++ + .. .H.
+
+ ---.-
----. ---
a
a
a
a
o.
a:C)
IT
o
z:
ITa
.-~
(f)-
I
I
(La
C)
~I~
WI
.-
0:
IT
LO
a
.
en
10.00
35.00
40.00
45.00
10.00
15.00
lIMEIOrS)
20.00 25.00 30.00
FROM STAAl OF TEST
5.00
Figure 19.
Martek IV pH vs. time. Data Set III.
-------
o
0
.
C\J
-
0 pHHL = -.053 + .937pHr /
o
-. /
- /
/
:]:0 /
CL~ /
0
m-
a:
-1
0
a:O +
°0
>- .
:]:Oi
/
/
0 /
0
.
CC /
. . +/{.
+ /
/
8.00 9.00 10.00 11. 00 12.00
STRNDRRD PH
o
o
.
f'-7 . 0 0
Figure 20. Hydrolab pH vs. standard pH, Data Set III. (The dashed
lines are the 95 percent confidence interval for the prediction of
mean pH values for the Hydrolab.)
obtained for actual pH values of 7 and 10:
Actual
7
6.5
7.1
8.2
pH
10
9.3
805
9.8
Hydrolab
Martek (MKIV)
Ocean Data
37
-------
r
o
t/)
.
N
--- .-.
~ .t if-
+
+ ~ + +.
ft T
+ .... + +
+ + t +
+ + ~+
tIJ..t "\-+ + + + + +
+ +
+ + +
"t.-*-It -I: .t It -f; + if + .... +
+ + + + + + + +
+ +
+
-
~.oo
10.00
15.00
T I ME (0)'5)
20.00 25.00 30.00
FROM STRRT DF TEST
35.00
40.00
45.00
o
o~
a:-
a:
0
z
a:
'-0
(f)~
o
I
(.oJ :c
en (Lo
LO
a: .
.-0
I
a:
0
20
CLt/)
W .
U-
01
o
t/)
.
N
.0.00
Figure 21. Hydrolab pH vs. time. Data Set III.
-------
After the field tests, the probes were cleaned and tested in the
laboratory with the following results.,
Hydrolab
(after cleaning)
r~a rt ek
(cleaning not required)
Ocean Data
(after cleaning)
These data suggest that the Hydrolab pH sensor did not drift significantly
over the range of calibration since the upper and lower values are all within
0.5 pH units of each other. If any drift did occur it was probably at the
lower set point. The Martek lower end point remained fairly constant, but the
upper end appeared to be affected by time from initial calibration since
biofou1ing of the Martek sensors was effectively eliminated. The Ocean Data
upper end point was fairly stable, the upper end increasing after cleaning,
but the lower end point seemed to be affected by fouling (increasing from 7.3
to 8.2) but recovered after cleaning. Thus the OD seems to be fairly stable
if kept clean, but is affected by fouling.
7.1
8.9
7.4
8.5
7.3
10.6
CONDUCTIVITY
Data analysis for the conductivity sensors from Data Sets I, II and III is
presented in Table 3 and plots from Data Set III are given in Figures 22
through 25. Specific observations are:
1. The conductivity measurements were by ~ar the most erratic of the
parameters measured. '
2. No conductivity measurements for the Ocean Data were obtained because of
an inability to calibrate the sensor.
3. Slopes ranged from -0.698 to 1.112.
4. Coefficients of Determination ranged from 0.01 to 0.75.
5. The 95 percent Confidence Intervals of the slope ranged from tl0 to t120
in fresh water, t1125 in brackish, t1300 in seawater.
6. The Hydro1ab tended to saturate at the lower end of the conductivity range
in the brackish and seawater data sets. For the range of conductivities
in which a linear trend was evident, the HL slopes were 0.531 and 0.556
with 95 percent Confidence Intervals equal to 10 percent and 3 percent'of
the average conductivity respectively for brackish and seawater.
7. In freshwater, the uncertainty ranged from 15 to 24 to 93 percent of the
average conductivity for the Hydro1ab.
8. The Martek uncertainty in freshwater ranged from 5 to 19 to 37 percent.
9. The significant deviation of the slopes from 1.0 suggests a similar
situation as was hypothesized for the pH sensors. No attempt to further
investigate this parameter was made due to time limitations.
39
-------
,-
TABLE 3. LEAST SQUARES REGRESSION ANALYSIS OF OBSERVED
CONDUCTIVITY AGAINST A DEFINED STANDARD CONDUCTIVITY
Summary of Least Squares Fit (LSF)
Data Range of
. Source Conductivity* 51 ope Intercept* R2t 0.95 CI* NDP!i
I, HL
freshwater 550-1000 1.021 -20 0.71 :t120 69
I, HLlB #
brackish 5500-8500 0.071 4679 0004 t600 28
I, HL2B#
brackish 8500-14000 0.531 922 0.55 t1125 50
I, HLlS**
seawater 31500-38000 0.251 24414 0.12 t3600 21
I, HL2S**
seawater 38000-45000 0 . 5 56 13394 0.75 t1300 41
I I, HL
freshwater 95-121 -0.698 198 0.01 tlOO 36
II, MK
freshwater 95-121 1. 7 99 137 0.49 t40 36
I II, HL
freshwater 150-270 0.767 20 0.37 t50 96
III, MKlt+
freshwater 150-270 00241 53 0.15 tlO 42
II I, MK2 t t
freshwater 150-270 1.112 -13 0.64 t40 41
* Conductivities in micromho/cm
t Coefficient of Determination, R2
* 95 percent Confidence Interval of the slope
~ Number of Data Points
. # The conductivity as indicated by the sensor was saturated over the
indicated range. The HL1B data are for standard conductivities in the
range of 5500 to 8500 and the HL2B data are for standard conductivities
in the range 8500 to 14000
** As for the brackish water data, the HLIS data are for conductivities in the
range 31500 to 38000, and the HL2S data are for conductivities in the range
38000 to 45000. .
tt The Martek package was removed after fourteen days exposure.. The MKI data
is for the first fourteen days data. After checking the dissolved oxygen
circuitry, the conductivity circuit was reca1ibrated and the package
redeployed. The conductivity as indicated by the sensor was very erratic
for a period of 5 days, then settled down. The MK2 data is for the period
after the sensor stabilized and the end of the test.
40.
-------
- ------- r------, ... -- ------- --- -- .._- _.- - -------
CM - -13. + 1.112Cr
- +
+
._- ~--------- - .-- -..- .----.-
+
+
.
++ + +
I 1111 I II I II I
+ II 11911*+
--
+
+
+
+
--+Hil+HHf--t----. ...-----
a
a
a
Ln
(T)
a
a
.
a
a
(T)
L
U
'A
La
La
--Ln
C\J
o
z
00
Uo
.
a
::::Co
wC\J
~
. a:
ITa
La
.
a
Ln
-
a
a
.
a
a
-100. 00
150+00 ++~OO. 00
STANDRRD CO NO
.,
250.00 300.00
( M t'"1 / C M )
350.00
Figure 22. Martek conductivity vs. standard conductivity,
Data Set II I .
41
-------
- -- ---
+
.
. -I-
+
+
+
.
+
++ +
+ + + -q.
+ ~ + + * .. '+
. ... .+ ..J.
+ T-ift- + iT+
+
+
. rt-/-t.t:.. +
-/r+.j!t- *1t. ~ +
- , - -j-- ---- -------
+
"-
I
o
o
.
o
::t'
Do
~-t 0
:.:.
o
(Y)
:E
U
,0
:EO
:EO
~(\J
.f.:o>
N
D
(L
ITa
00
Zo
IT~
r-
U1
I
o
00
Zo
£)
U
a
>0
t--t
o
~I
W
r-
(La
ITa
:E
a
C\J
'0.00
5.00
10.00
Figure 23.
IS.OO
TIME lOYS)
20.00 25.00 30.00
FROM START OF TEST
35.00
Martek conduct i vity vs. time. Oata Set II I.
40.00
45.00
-------
a
a
.
c
,,",
{r;;
!
a! Cn = 20 + .767C,-
a!
. i
a:
a'
(T")I
..-. I
I
I
I
i
I
i
;
I
i
i
I
i
,it
/'
/
i/
/
t /" i
I ~ I
I ~
:/
V
/i
i
j
a!
a!
. I
a.
a!
-100.00
!
, ;
!
,
;
/
/
V
./
/'
..
, ,
, .
i .
I !
I '1
I ~ ~
i ./ I
!/
~
u !
..........a;
~a :
~oi
-oJ) .
~ '
:
o
z
Oa:
( .,-., .
'-J-. i
a,
C!J,=. '
~('~ : 1,/
r., ~.
- I , I
C::a: /' t-
00: /' i
>-0; ./ '
::r:..D I
/:
/ + j
1./ !
~ i '
/; ,
+
./
+ +' ./
+ .. .v
/,
ISO.OO 200.00
STR~ORR:J [aND
, I
250.00 300.00
(MM/CMJ
I
350.00
Figure 24. Hydrolab conductivity vs. standard conductivity, Data
Set III. (The dashed lines are the 95 percent confidence interval
for the prediction of mean conductivity values for the Hydrolab.)
DISSOLVED OXYGEN
Table 4 presents an analysis of data from Data Sets I, II, and III and for
DO measurements, and Figures 26 through 29 are plots of the data obtained in
Data Set III. Observations about the data are as follows.
10 The Hydrolab and Ocean Data DO sensors performed well in that the slopes
ranged from 0.954 to 1.073 and Coefficients of Determination ranged from
0.88 to 0.97.
43
-------
.,:a
~
o
o
.
o
o
-
+
-1-
t t it- -t- +
-1*++ ft +
+++* + + + + # .t- + *++ +
if- ++ -fI- ++ jt ++ + ,...p- * + + ++ t
+ + + + + -
+ -t +. + + +
+ 't +
-- I
+
+
"
+
.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 Ll t 00 11
~
Uo
"'-0
~ .
~o
~lJ)
o
a:o
0:0
o.
ZO
0:
I-
<.no
o
I .
o
lJ)
O'
Z
00
Uo
.
(Do
cr:~
--' ,
o
a:
og
>- .
IO
lJ)
'0
5.00
TIME(O)'S)
FROM STRRT OF TEST
Figure 25.
Hydrolab conductivity vs. time, Data Set III.
-------
TABLE 4. LEAST SQUARES REGRESSION ANALYSIS OF ORSERVED DISSOLVED
OXYGEN AGAINST A DEFINED STANDARD DISSOLVED OXYGEN
Summary of Least Squares Fit (LSF)
Data Ran{e of Interce)t R2....
Source DO ppm) Slope (DO ppm 0.95 CIt MDP*
I, HU
fres hwate r 2-12 0.954 0.105 0.95 :t1.15 37
I, HU
brackish 3-9 1.073 -0.411 0.88 :t1.05 37
I, HL 9
seawater 5-9 1.051 -0.490 0.89 :to.80 30
I I, HL
freshwater 7-18 0.961 0.735 0.90 ::1.70 35
II, MK
freshwater 7-18 0.666 4.782 0.62 :t2.75 36
II I, HL
freshwater 0-20 0.966 -0.296 0.97 :t1.95 74
III, 00#
freshwater 0-17 1.048 -0.124 0.96 :t2.10 67
... Coefficient of Determination, R2
t 95 percent Confidence Interval of the slope
* Number of Data Points
S The circulator failed before the test run was completed. The data
presented were taken before the ci rcul ator fail ed.
# The Ocean Data DO sensor saturated. at DO greater than 17 ppm. The data
presented are for those values of standard DO less than 17 ppm.
45.
-------
o
a
.
!.J')
C'\.i.
!
'A
I DOoD= - .124 + 1.048DOr
o.
o!
. !
0:
C\.;
!
- .
I
I
I
I
!/
/f + ~ +
I
A ++ +.
i ,L ~ +.: / .
! ~. +..,..:~
: / -+- 4:- +/
")tt ...+ .F + / .
1+ '. /
~ +:~ /
~ : /
'*Y
+ / i
A-
+
+~
+
+
,....,.
~
Q...
a.... a
_0.
. . ~
. 1.f') :
0-,
o .
c:
i--O :
~o :
...... .
I
i
I
. i
Co.
I
! /
j/
L..:..J ~
; ),
U i /1,+
00, +
0; / i
. .
1.f')j //. ! /
: y
V + /!
I
0: + /
0' ...
! . '
.
z
c
.;,
i
;
O'
o.~o
5'.00 1'0.00
STRNDRRD 00
,
15.00
(PPMJ
I
20.00
25.00
Figure 26. Ocean Data DO vs. standard DO, Data Set III. (Values
above 15.4 DO units on the x axis were not used in detennining
the regression equation. The dashed lines are the 95 percent
confidence interval for the prediction of mean DO values for
the Ocean Data.) .
46
-------
a
a
.
::J'
t-
(f)
EJ
~Eb
L~
eLru
eL
0
0:0
aD
00
Z
a:
~ t-
....... (f)
0
o
I .
ru
I
0
0
aD
t-o
a .
o:;:r
I
+
++ +
+ + +
"1" + +
+
+ + t- ...
+ + + :t + +
+ + +
+.
+ + +
+ + .1. + ff ....
"1" + + t t -tI-' -I- +
+ -1-+ ... + t-l- + +ij.+
t- +
+
+
+ +
+ +
+
+
+
+
_._---~
z
a:
w
UO
o
o .
to
'0.00
5.00
10.00
15.00
TIME (OYS)
Figure 27.
20.00 25.00 30.00
FROM START OF TEST
35.00
Ocean Data DO ys. time, Data Set III.
40.00
45.00
-------
o
o
.
lJ)
('\,;
!
i
o!
0'
I
. I
I
O.
(\,;
-- I
~ j
Q. i
Q.o I
O.
- I
lJ)
0-:
Q :
co i
r-O .
---;0:
001
cc- ;
o '
>-
-r-
o ~
0:
A
DOHL= -.296 + .966DOr
I I
i
I
. .
.
.
lJ):
I /
1/
~
0,
~I
00.00
- I
; I
5.00 10.00
STRNDARD 00
15.00
(PPM)
I ~ !
1/ I
r /1
I
I
+i /
V
/:
I
I
20.00
I
25.00 .
Figure 28. Hydrolab DO vs. standard DO, Data Set III. (The
dashed lines are the 95 percent confidence interval for the
prediction of mean DO values for the Hydrolabo)
48
-------
o
0-
.
;:jI
:E
(Lo
(Lo
'-' .
(\J
o
cr:
0:0
00
Z .
0:0
I--
(j)
~
U) 10
o
.
(\J
O'
o
(00
0:0
--.J~
01
cr:
0
>-
Io
o
'_n
+
++ + -
+ + + +
++ + +
+
fr+ + + + + ..,..... + ...
+ + + ft*
+ + + + +
+ ++t- + + +
++ .. + +
+ + + +
+ +
+ + + T :r
+
+
-
.
(£)
'0.00
10.00
15.00
TIME (OYS)
20.00 25.00 30.00
FROM STRRT OF TEST
5.00
ngure 29.
Hydrolab DO vs. time. Data Set III.
35.00
110.00
115.00
-------
~
2. All slopes for the HL and 00 were not statistically different from 1.0 at
the 5 percent probability level.. .
3. The 95 percent Confi dence Interval .for the HL ranged from :t1.05 to :t1.95.
4. The 95 percent Confidence Interval for the 00 was :t2.10 for 00 values from
o to 17. At values of 00 greater than 17, the 00 sensor appeared to
saturate and level out. .
5. The Martek 00 sensor gave very erratic data for the the 41-day test and
was not deemed reliable enough to perform an analysis.
6. The Martek data for Data Set II give a slope of 0.666, coefficient of
determination of 0.62 and an uncertainty of t2.75. This slope is
statistically less than 1.0.
7. The MartekOO sensor after 41 days exposure, protected by the enclosure
and antirouling paint, gave a slope of 0.683, R2 of 0.943 and an
uncertainty of t1.4 ppm (Data Set IV).
CAL IBRA TI ON
. The instruments were cal i brated accordi ng to the manufacturer IS.
instructions. The calibration procedures call for calibration setpoints at
most, 2 points, preferably over the range of parameter values expected in the
fiel d measurements. ..
In the 6-week field test discussed in this report, the pH range from 7.5
to 11 and the dissolved oxygen (DO) ranged from 0 ppm to 20 ppm. Due to the
very 1 arge range of va ri at i on of the pH and DO, the cali brat i on procedures
should be changed to allow additional setpoints. This would require changes
in the electronic circuitry to permit additional setpoints to be added. This
would result in greater stability over the entire range of parameter
variation. Therefore, the drift of anyone setpoint would not affect the
entire calibration curve.
Due to the inher~nt variation of the probe readings, it would be desirable
to take several probe measurements of a given parameter over a period of time
after the initial calibration, holding the actual parameter value constant.
The average of the resultant distribution of probe readings should then agree
with the actual value of the parameter being measured. If it does not, then
the difference between the actual value and the probe average reading should
be added to the current probe reading and the circuitry readjusted to reflect
the new calibration. This procedure may be repeated if necessary until the
average of the indicated probe readings is equal to the actual value of the
parameter being measured. This procedure would eliminate any calibration bias
due to a one-time calibration adjustment.
50
-------
SUMMARY OF DATA FOR ALL PARAMETERS
Table 5 presents a summary of statistical tests for null hypothesis for
each parameter and each data set. Whether or not the data were statistically
acceptable is shown.
Table 6 presents the results obtained from Data Set IV, the data taken
after field tests were completed and the probes cleaned. The tests were
accomplished in the instrumentation laboratory at LSU under controlled
conditions.
51 .
-------
TABLE 5. SUMMARY OF STATISTICAL TESTS FOR NUll HYPOTHESIS
(Ho: b = 1)
Data
Source Parameter b* SO of bt t:f: AcceptS Rejects
I, Hl
freshwater T 0.934 0.059 -1.12 *
I, Hl
brackish T 0.991 0.089 -0.10 *
I, Hl
seawater T 1.065 0 .025 2.60 *
I I, Hl
freshwater T 0.995 0.048 0.10 *
II, MK
freshwater T 1.013 0 .026 0.50 *
I II, Hl
freshwater T 1 0017 0.018 0.94 *
II I, MK*
freshwater T 0.955 00025 -1.80 ...
III, MKt
freshwater T 0.902 0.018 -5.40 *
II I, 00
freshwater T 0.986 0.033 -0.42 *
II, Hl
freshwater pH 0.503 0.061 -8.15 *
II, MK
freshwater pH 0.666 0.061 - 5.50 ...
I II, HL *
freshwater pH 0.807 0.058 -3.33 ...
I II, MK*
freshwater pH 0.732 0.079 -3.40 *
III, MKt
freshwater pH 0.781 0.145 -1.51 *
(continued)
52,'.
-------
TABLE S. (Continued)
Data
Source Parameter b* SD of bt t* AcceptS RejectS
-III, MK*
fres hwater pH 0.657 0.016 -21.4 *
III, MKS
freshwater pH 0.464 0.028 -19.1 *
I II, 00*
freshwater pH 0.793 0.067 -3.09 *
II I, 00
freshwater pH 0 .525 0.035 -13.47 *
I II, HL t
freshwater pH 0.937 0.040 -1.58 *
I, HL
freshwater C 1.021 0.079 0.27 *
I, HL2B
brackish C 0.531 0.070 -6.70 *
I, HL2S
seawater C 0.556 0.052 -8.54 *
II, HL
freshwater C -0.698 1.360 1.25 *
II, MK
freshwater C 1.799 0.550 1.45 *
I II, HL
freshwater C 0.767 0.100 -2.26 *
I II, MK*
freshwater C 0.241 0.093 -8.16 *
III, MKt
freshwater C 1.112 0 .134 - 0.84 *
(continued)
53
')
-------
TABLE 5. (Continued)
Oata
Source Parameter b* SO of b t. t* Accept. 9 Rejects
I, HL
freshwater 00 00954 0.039 -1.18 *
I, HL.
bracki sh 00 10073 00068 1.07 *
I, HL
seawater 00 10051 00070 0.71 *
II, HL
freshwater 00 0.961 00057 -0068 *
II, MK
freshwater 00 0.666 0.090 -3.71 *
II I, HL
freshwater 00 0 .966 . 0.019 -1.70 *
I II, 00
freshwater 00 1.048 0.027 1.78 *
* b = Slope of Least Squares Fitted curve
t SO of b = Standard deviation of slope
* t = b - 1.0
SO of 8
s Accept or reject at 0.05 probability level
54
-------
-- -""~_. - .j - - ~--_....L"---
TABLE 6. DATA SET IV. CALIBRATION CHECK OF PROBES AFTER CLEANING
Probe Parameter Slope Intercept SO of b t Accept* Reject* 0.95 CI NOP
HL 00 0.810 0.477 0.045 4.22 * 0.9 17
00 00 0.928 0.927 0.052 1.38 * 1.0 17
MK 00 0.683 0.030 0.085 3.73 * 1.4 10
In
In
HL pH 0.615 2.742 0.113 3.407 * 0.3 17
00 pH 0.809 1.076 . 0 .140 1.364 * 0.3 16
MK pH 0.381 4.689 0.093 6.656 * 0.2 15
* Accept or reject at 0.05 probability level.
(Temperature and conductivity remained essentially constant during this laboratory test.)
-------
SECTION 7
REFERENCES
American Public Health Association. 1976. Standard Method for the
Examination of Water and Wastewater, 14th Edition.
Anderson, D. B., and B. R. Richards. 1966. "Chlorination of Sea Water
Effects on Fouling and Corrosion," Journal of Engineering for Power, Vol.
88. pp. 203-208.
Best, G. A. 1974. "Continuous Water Quality Monitoring." Effluent and Water
Treatment Journal. 14(7):345-362. .
Caprari, J. J., and V. Rascio. October, 1972. "Study of Some Variables
Affecting Anti-fouling Paints' Performance," Third International Congress
on Marine Corrosion and Fouling. Gaithersburg, Maryland. pp. 938-939.
Connolly, Brian. 1973. "In-situ Hypochlorite Combats Marine Growths in
Blocked Pipework," Process Engineering. pp.96-97. .
Cooke, G. H., and G. M. Woodward. 1970. "tontinuous Automatic River Water
Quality Monitoring - Experience on River Trent." Effluent and Water
Treatment Journal. Vol. 10, No.5. pp.257-264. .
Davies, A. W. 1972. Water Quality Monitoring in Europe.. Water Pollution
Control '71. Vol. 3. pp. 299-305.
Dyckman; E. J., E. C. Fischer, and J. A. Montemarano. 1974. "Antifouling
Organometall i c Structural Pl asti cs, II Naval Engi neers Journal. Vol. 86,
No.2. pp. 59-64.
Eller, S. A., R. G. Grunther, and Edel Stein. 1970. "Fouling Resistant.
Elastomeric Material for Sonar Domes of Naval Surface Vessels," Naval
Engineers Journal, Vol. 82, No.1. pp.115-121.
Gasperecz, G. J. 1978. An Evaluation of Probe Performance and Remote Water
Quality Monitoring in Coastal Louisiana. Master of Engineering Thesis,
Louisiana State University, Baton Rouge, Louisiana. pp. 1-69.
Hydrolab, Inc. 1975. Instructions for Operating the Hydrolab Surveyor Model
60 In-situ Water Quality Analyzer. pp. 1-148.
56
-------
Klein, W. L., D. A. Dunsmore, and R. K. Horton. 1968. IIAn Integrated'
i40nitoring System for Water Quality Measurement in the Ohio Valley.1I
Presented at the National Meeting of the American Chemical Society
Symposium on Instrumental and Automated Methods of Chemical Analysis
Water Pollution Control. San Francisco, California.
for'
LaBarge, Inc. April, 1976. Instruction Manual for the Convertible Data
Collection Platform (CDC?) and Related Equipment. Tulsa, Oklahoma.
Lamb, Thomas J. October, 1972. IIMarine Fouling Control by Electrolytic
Hypochlorite Generation,1I Third International Congress on Marine Corrosion
and Fouling. Gaithersburg, Maryland. pp. 995-1004.
Montemarano, J. A., and E..J. Dyckman. 1912. Antislime Coatings Part III:
Antislime Organometallic Polymers of Optical Quality. Research and
Development Report 3597. Naval Ship Research and Development Center,
Bethesda, Maryland.
Montemarano, J. A., and S. A. Cohen. 1976. Antifouling Glass-Reinforced
Composite Material. Research and Development Report MAT-75-33. Naval
Ship Research and Development Center, Bethesda, Maryland.
Palmer, M.D., and J. B. Izatt. 1970. Determination of Some Chemical and
Physical Relationships from Recording Meters in Lakes. . Water Research,
4 ( 12) : 773- 786.
Rohling, Joseph P., William B. Waff, and Paul A. Wolfgram. 1975.
IIDevelopment of a Water Qual i ty I nstrumentati on Package for Long-Term
Operation from Buoys and Other Unattended r4arine Platforms.1I In:
Proceedings of the Seventh,Annual Offshore Technology Conference, Houston,
Texas. . pp. 529-536.. .
Saroyan, John R. 1968.
. Engineers Journal.
"Antifouling Paints - The Fouling Problem.1I Naval
Vol. 80, No.4. pp. 593-604.
Smith, B. L., Jr. 1973. Design .and Development of an Estuarine Monitoring
Buoy System for Prediction Moaelolnvestigation. Ph.D. Thesis, University
of Rhode Island, Kingston, R!r. Vol. I. 231 pp.
. .
West, H. W., and H. M. Floyd. 1976. An Automated System for Collecting,
Processing, and Displaying Environmental Baseline Data. Technical Report
M-76-11. U.S. Army Engineer Waterways Experiment Station, Vicksburg,
Mississippi. 88 pp. .
Wruble, D. T., J. D. Koutsandreas, and B. Bijonowski. 1978. In:
Proceedings of the Automated In-situ Water Quality Sensor Workshop. U.S.
Environmental Protection Agency, Las Vegas: Nevada. pp. 1-5.
57
-------
TECHNICAL REPORT DATA
(Please read InstrUctions on the reverse before completing)
,. REPORT NO. 12. 3. RECIPIENT'S ACCESSION NO.
r:PA-nOn; 4-BO-052
4. TITLE AND SUBTITLE 5. REPORT DATE
THE DEVELOPMENT OF A DEPLOYABLE WATER QUALITY November 1980
MONITORING SYSTEM 6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S) 8. PERFORMING ORGANIZATION REPORT NO.
C. A. Whitehurst and G. D. Whitehouse
9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT NO.
Division of Engineering Research 1BD884
College of Engineering 11. CONTRACT/GRANT NO.
Louisiana State University >
Baton Rouge, LA 70803 Grant No. R806313010
12. SPONSORING AGENCY NAME AND ADDRESS 13. TYPE OF REPORT AND PERIOD COVERED
U.S. Environmental Protection Agency--Las Vegas, NV Prnipl"t Rnt. 1n/1/7Q_nO/1.n/70
Office of Research and Development 14. SPONSORING AGENCY CODE
Envi ronmenta1 Monitoring Systems Laboratory
Las Vegas, Nevada 89114 EPA/600/07
15. SUPPLEMENTARY NOTES
16: ABSTRACT
The stability and accuracy of current off-the-shelf water quality sensors are
evaluated. Biological fouling was deemed one of the most serious problems in
nullifying sensor capability. Methods to reduce biofou1ing were evaluated. Of the
potenti a1 measures to eliminate biofouling, the focal "kil1-zonell approach appears to
be most feasible and the simplest to implement.
-.
17. KEY WORDS AND DOCUMENT ANAL YSIS
a. DESCRIPTORS b.IDENTIFIERS/OPEN ENDED TERMS c. COSATI Field/Group
Water poll ution 13B
")
18. DISTRIBUTION STATEMENT 19. SECURITY CLASS (This Report) 21. NO. OF PAGES
UNCLASSIFIED 65
RELEASE TO PUBLIC 20'~Et~S"Slf:1rlj (This page) 22. PRICE
i
i'
SPA Form 2220-1 (Re", 4-(7)
PREVIOUS EOIT10N IS OBSOL.ETE
------- |