Content uploaded by Michael Exner-Kittridge
Author content
All content in this area was uploaded by Michael Exner-Kittridge on Nov 17, 2014
Content may be subject to copyright.
A simple and flexible field-tested device for housing water
monitoring sensors at point discharges
Michael Exner-Kittridge, Richard Niederreiter, Alexander Eder
and Matthias Zessner
ABSTRACT
The Water Monitoring Enclosure (WME) provides a simple and flexible housing for many types of
sensors for continuous measurements of water parameters (physical, chemical, or biological) and
provides the opportunity of representative sampling for external analyses. The WME ensures a
minimum internal water level and this ensures that the internal monitoring equipment remains
submerged even when there is no flow into the enclosure. The limited diameter of the inflow pipe
and water volume in the WME buffers the flow velocity from dramatic changes. The device
ensures that the sediment entering the enclosure from the inflow will be conveyed through the
enclosure with minimal sediment accumulation. The device is powered purely from natural hydraulic
forces, so it requires no power source, and requires little additional maintenance beyond periodic
cleaning. If desired, the WME can also measure discharge entering the device through additional
modifications. Water samples were taken throughout the year to validate the effectiveness of
the WME. The comparisons of the influent water to the water in the WME for all parameters were
below the laboratory analysis standard error or below the limit of quantification, indicating that
the water in the WME is representative of the influent water.
Michael Exner-Kittridge (corresponding author)
Vienna University of Technology,
Centre for Water Resource Systems,
Karlsplatz 13/222, A-1040,
Vienna,
Austria
E-mail: mgkittridge@gmail.com
Richard Niederreiter
UWITEC, Weißensteinstraße 305310,
Mondsee,
Austria
Alexander Eder
Federal Agency for Water Management,
Institute for Land and Water Management
Research,
Pollnbergstraße 1,
3252, Petzenkirchen,
Austria
Matthias Zessner
Vienna University of Technology,
Institute of Water Quality,
Resources and Waste Management,
Karlsplatz 13/222, A-1040,
Vienna,
Austria
Key words |environmental monitoring, sensor housing, water monitoring, water quality
INTRODUCTION
Measurements and samples of water parameters are a funda-
mental aspect of all water-related research and monitoring
aside from purely theoretical research. As technology
improves and scientific questions plunge deeper into the
finer scales of time and space, water measurements have
been migrating from grab measurements and samples
taken personally at a field site to permanently or temporarily
installed instrumentation continuously collecting data. This
type of instrumentation has clear advantages. Researchers
can acquire data at a high frequency and during periods or
at locations that might be difficult to capture (i.e. short dur-
ation rainfall events, distant or temporarily inaccessible field
sites, etc.). Unfortunately, continuous monitoring equipment
also tends to be much more complex and subsequently more
maintenance-prone than simple grab measurements or
samples.
Direct laboratory analysis is still the most accurate
method to analyze water chemistry as the equipment is
regularly maintained, calibrated, and validated to ensure
a specific level of data quality. Researchers who only
require a couple of dozen samples for periodic field tests
have used automatic samplers to take samples remotely
at a high temporal frequency (Tan et al.;Burns
et al.;Roser et al.;Ensign & Paerl ). But
for those who require a high temporal resolution in
addition to longer monitoring programs spanning months
or years, the costs for both the laboratory analyses and
labor become prohibitive.
Researchers have recognized that if long-duration high
temporal resolution measurements of water chemistry are
to be performed then the laboratory would need to be
mostly divorced from the actual measurement process. Cur-
rently, there are two overarching methods to remotely
collect high temporal resolution water chemistry using tem-
porary or permanent installations. The first is typically
known as on-line measurement devices. These devices
1026 © IWA Publishing 2013 Water Science & Technology |67.5 |2013
doi: 10.2166/wst.2013.655
effectively miniaturize and automate aspects of the labora-
tory analysis (i.e. filtration, addition of chemical additives,
etc.) at a fixed location directly at the water body for
sampling. The second is to use an in situ sensor as a proxy
for the measurement of the chemical compound. The
former method has been used extensively in wastewater
monitoring (Legnerová et al.;Carrasco et al.;Sten-
holm et al.). The ability of an automated on-line device
to draw in a sample and prepare the sample through
filtration or chemical additives is useful when the sample
may have high concentrations of many toxic substances.
The primary trade-off of such a highly complex device
with automated pumping and filtration systems is
that it requires a great deal of maintenance to upkeep the
device.
In situ sensors have become increasingly popular within
the last couple decades for both wastewater and non-waste-
water applications (Newman ;Kaelin et al.;Kim
et al.;Kestel et al.). Hydrologists have been
using similar devices for many decades to measure physical
water parameters (i.e. water level, temperature, electrical
conductivity, etc.). The advantages of in situ sensors include
a very small size relative to the on-line devices, simpler con-
struction and operation, lower cost, and temporal resolution
that can be finer than on-line devices. The primary disadvan-
tage is a lower accuracy and higher detection limit to the on-
line devices.
Both the on-line devices and the in situ sensors have
physical measurement constraints that must be taken into
account before installation to ensure proper measurement
quality. The constraints become especially problematic
when measuring in natural environments with low dis-
charges (<5 l/s). Both types of devices require a water
depth of at least 20–30 cm depending on the type of
measurement device. The water could either be a free-
flowing stream or an intermediate pool of water. For sen-
sors with polymer membranes, the water level must not
drop below the measurement device as the sensors must
always stay wet. Additionally for in situ sensors, large
ranges in the water velocity can significantly affect elec-
trochemical signals. If parameters associated with
suspended solids are to be measured (i.e. turbidity, phos-
phate, etc.), then resuspension during storm flow events
should be greatly limited to ensure that the sampling is
representative of the influent water. Suspended solids
sampling is generally more difficult to be representative
of the influent water than dissolved solids sampling.
Harmel et al.()found that, while sampling uncer-
tainty for dissolved solids was ±25%, the uncertainty for
suspended solids was over 50%. Finally, continuously
deposited sediments into the measurement pool could
reduce the optimal water depth and in the worst case
bury the measurement device.
Given the above constraints for the installation con-
ditions for water measurement devices, researchers
traditionally have had two general options when installing
equipment in natural environments. The first is to simply
excavate a pool directly below the influent water and
place the measurement devices (or pump tubing) directly
into the pool. The main drawback of the pool is that it
requires frequent excavation of the sediment deposited
into the pool. Due to the reduction in the flow velocity
from the influent water to the pool, sediment will inherently
deposit into the pool and not be naturally transmitted
through the system. Additionally, the flow velocity in the
pool could vary dramatically depending on the volatility of
the influent water. The second option is also to excavate a
pool, albeit a smaller one to that of the first option, and
use a tube with a pump to draw up the water into an external
container where the measurement devices are housed. The
pump–container option is advantageous in that the exca-
vated pool can be much smaller than for the first option,
and subsequently the sediment deposition is reduced. The
pump –container option would ensure both a minimum
water level and a consistent flow velocity for the measure-
ment devices. The key disadvantage for the pump–
container option is the pump itself. Additional mechanical
moving parts like a pumping system increase the likelihood
of malfunction and breakdown especially during winter
months where freezing can be an issue. An ideal option
would have the relative simplicity of the first option, the con-
sistent flow velocity of the pump–container option, and the
ability to continuously transmit sediment through the
measurement system.
We have designed, built, and tested a measurement
system with all of the above constraints. The Water Moni-
toring Enclosure (WME) provides a simple and flexible
housing for sensors for continuous measurements of
many types of water parameters (physical, chemical, or bio-
logical) and provides the opportunity of representative
sampling for external analyses (e.g. remote mini-labora-
tories or automatic samplers). The WME ensures a
minimum internal water level and this ensures that the
internal monitoring equipment remains submerged even
when there is no flow into the enclosure. The limited diam-
eter of the inflow pipe and water volume in the WME
buffers the flow velocity from dramatic changes. The
device ensures that the sediment entering the enclosure
1027 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
from the inflow will be conveyed through the enclosure
with minimal sediment accumulation. The device is pow-
ered purely from natural hydraulic forces, so it requires
no power source, and requires little additional maintenance
beyond periodic cleaning. If desired, the WME can also
measure discharge entering the device through additional
modifications.
TEST SITE
The test site chosen for the WME was the Hydrologic Open
Air Laboratory (HOAL) catchment located in Petzen-
kirchen in Lower Austria, approximately 100 km west of
Vienna. The catchment has an area of 67 hectares and the
land cover is characterized as 90% agriculture, 5% imperme-
able surface, and 3% forest. There are 12 point discharges
that drain into the catchment stream. These include seven
subsurface tile drains, two springs, and four surface tribu-
taries. Baseflow can range from 0.00 l/s at some of the
subsurface drains and surface tributaries to 5 l/s at the catch-
ment outlet. Average baseflow at the subsurface drains and
surface tributaries is 0.1–0.2 l/s with a maximum measured
discharge of approximately 13 l/s. In total, four WMEs
were installed within the catchment.
DESIGN AND CONSTRUCTION
Figure 1 illustrates the individual components of the WME.
There are five primary components: (1) center shaft, (2)
plunger, (3) bladder, (4) outer enclosure, and (5) outflow
siphon.
Figure 2 represents the fully constructed WME. The
WME is almost entirely composed of polyvinyl chloride
(PVC). The only exception is the plunger rod, which is com-
posed of stainless steel. This type of PVC was primarily used
because of its low corrodibility and ultraviolet resistivity.
The inflow piping to the WME is standard 50 mm diameter
PVC attached directly to the WME. The overflow should be
located directly opposite the inflow with a 50 mm diameter
pipe. The outflow opening at the bottom of the center shaft
is 40–45 mm diameter. A 75 mm diameter PVC pipe is sub-
sequently attached to the bottom of the center shaft, which
is ultimately reduced to 50 mm once it connects to the
siphon. The plunger is set within the center shaft with the
bladder set around the center shaft. Both the plunger and
Figure 1 |The individual components of the WME and associated dimensions: (1) center shaft, (2) plunger, (3) bladder, (4) outer enclosure, and (5) outflow siphon.
1028 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
the bladder should be of a diameter that would allow them
to move freely within the center shaft with a changing water
level. An adjustable screw fixed to the upper end of the plun-
ger and set atop of the bladder ensures that when the
bladder moves vertically the plunger will also move with
it. The adjustable screw on the plunger rod can be raised
or lowered to modify the minimum and maximum water
level within the WME. The plunger should have a flexible
rubber seal on the bottom end to facilitate a watertight
seal when in the closed position. The total height of the
WME without the center shaft should be determined by
the instrumentation to be placed inside the WME. Our
WME had a wall height of 30 cm and a total height to the
bottom center of the WME (excluding the center shaft) of
35 cm. An inverted siphon should be placed at the outflow
piping of the WME and should not rise above the bottom
of the center shaft where the plunger would be in the
closed position. The overflow should be connected to the
outflow piping directly after the siphon using a ‘T’joint
PVC pipe as shown in Figure 2.
FUNCTION
The WME is designed to provide four key functions: (1)
maintain a minimum water level to ensure that the measure-
ment sensors do not become dry, (2) allow sediment to be
continuously transmitted through the WME, (3) limit the
flow velocity range, and (4) be simple and require little
maintenance.
The 50 mm diameter inflow restricts the amount of flow
into the WME that can be readily discharged at the bottom
of a 40–45 mm diameter outlet (Figure 1). The inflow restric-
tion removes the need for a large overflow in case of severe
clogging of the outflow and limits the magnitude of the flow
velocity in the WME. The inflowing water is directed per-
pendicularly into the WME to allow for proper circulation
and mixing within the WME. The sloped bottom funnel
ensures that any sediment deposition will be gradually trans-
ported down towards the WME outlet during flushing and
ultimately removed.
When a maximum water level threshold is reached
within the WME, the bladder quickly lifts the plunger to
allow a high-velocity flow pulse through the outflow pipe.
Following the initial opening of the plunger, the water
level gradually drops since the outflow discharge is greater
than the inflow discharge. Once the water level reaches a
minimum threshold, the bladder lowers and the plunger
closes the outflow. The mechanism is intended to always
fully open regardless of the inflow rate to ensure proper sedi-
ment conveyance. In some instances with a long outflow
pipe, the closing of the plunger to the outflow may bounce
several times before completely closing, owing to natural
water hammer.
The inverted siphon shape was configured to retain a
volume and level of water directly below the outflow, but
not as to exceed the height of the plunger inside the WME
as this can cause poor sealing of the plunger. The closed
volume of water in the siphon ensures that, when the
water level in the WME reaches the maximum threshold,
and the plunger is subsequently lifted by the bladder, the
water exiting the outflow does not exert a counter force on
the plunger potentially pulling the plunger back to a
closed position. Without the siphon, the bladder and plun-
ger mechanism would not open fully when the initial
maximum threshold is reached, and consequently a steady
state would be reached where the inflow would equal the
outflow. Due to the high velocities in the siphon compared
with the inflow and within the WME, no sediment accumu-
lation occurs in the siphon.
Discharge measurements
When water is regularly entering the WME, the plunger
opens and closes at a regular frequency relative to the
incoming flow. With a consistent flushing frequency from
the incoming flow, measuring the time between the flushings
allows for the estimation of discharge from the WME. The
mechanism functions similarly to a tipping-bucket rain
Figure 2 |Fully assembled and installed WME with foundation poles. The holes in the lid
of the WME are for the placement of the measurement sensors.
1029 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
gauge and can be installed with similar instrumentation.
Q¼t2t1
V(1)
where Qis the discharge, t
1
is the time of the initial opening
of the plunger, t
2
is the time when the plunger closes, and V
is the difference between the volume of water at the maxi-
mum water level and the minimum water level in the
WME. In this situation, Vis not equivalent to the volume
of flushed water as the flushed water includes the inflow
water during the flushing in addition to the V. For practical
purposes, it may not be possible to measure both an opening
time and a closing time, which could require two separate
devices logging two separate time stamps. Measuring the
time once per opening is more feasible, but a relationship
must be derived between the times of the sequential open-
ings of the plunger to the time between the opening and
the closing of the plunger.
FUNCTIONAL ASSESSMENT
Several tests were performed to verify the effectiveness of
the WME. The tests included (1) a verification that the
water chemistry in the WME is representative of the influent
water and (2) a verification of the regularity of the flushings
for use in estimating discharge.
Chemical assessment
For the chemical assessment, one sample was collected at
the influent water (directly before the intake of the WME)
and one sample was collected from within the WME at
four separate times over a 1-year period. Table 1 summarizes
the results from the chemical analysis of the samples. The
sample representativeness error was estimated with the fol-
lowing equation:
SRE ¼
Ca1
Cb1
1
þCa2
Cb2
1
þ...
n(2)
where SRE is the sample representativeness error, C
a
is the
concentration of the influent water, C
b
is the concentration
of the water in the WME, and nis the total number of
sample sets. In our comparison, we had four sample sets.
In addition to the four sample sets that were taken over
the course of a year, another set of 25 samples were taken on
1 day at the same WME to test the laboratory analytical
error. These samples were taken in quick succession and
with great diligence to ensure that minimal error could be
attributed to the sample collection and sample transpor-
tation. Our term laboratory analytical error includes both
sample preparation and analytical error. Only three chemi-
cal parameters (phosphate, nitrate, and chloride) were
analyzed due to cost limitations. The laboratory analytical
error was calculated by the coefficient of variation
(standard deviation divided by the mean) of all 25 samples.
The results are shown at the bottom of Table 1 as lab analyti-
cal error.
The majority of the parameters had neither a consistent
bias nor an error greater than 2% of the original measure-
ments. Although several parameters (e.g. dissolved organic
carbon (DOC), total organic carbon (TOC), phosphate,
ammonium, and total phosphorus) had larger errors, these
parameters were near or below the limit of quantification
of the laboratory analytical techniques. Phosphate was
close to the quantification limit of the laboratory analysis,
while nitrate and chloride were well above the quantifi-
cation limit. Nitrate and chloride with a concentration
well above the detection limit had errors of approximately
0.01 and 0.05 respectively, while phosphate with a concen-
tration near the detection limit had an error of
approximately 0.30. Indeed, the error found between the
influent water and the WME is well under the error for
phosphate and at or below the error for nitrate and chloride.
DOC and TOC had errors above 3%, and we were unfortu-
nately not able to derive laboratory-specific analytical errors.
According to the DIN/EN 1484 standards for the laboratory
analysis of DOC and TOC (Committee EH/3/2 ), the
analytical error can be on average 1–2 mg/l. As the average
differences for the sample sets for DOC and TOC were 0.13
and 0.35 mg/l respectively, the SRE can be attributed to the
analytical error. During a normal sampling campaign,
sample collection, sample transportation, sample prep-
aration, and sample analysis all add error. Harmel et al.
()found that under typical conditions the cumulative
probable uncertainty was 4–48% for sample collection, 2–
16% for sample preservation/storage, and 5–21% for labora-
tory analysis. For this reason, it is surprising that our sample
comparisons between the influent water and the WME had
only a couple percent error for most of the parameters.
Discharge assessment
The data for the assessment of the consistency of the flush-
ing frequency of the WME during constant discharge were
1030 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
Table 1 |Chemical comparisons between the influent water and the WME taken at four times throughout the year
Location Date
Dissoved
organic
carbon
(mg/L)
Total
organic
carbon
(mg/L)
Phosphate
(mg/l)
Ammonium
(mg/l)
Nitrate
(mg/l)
Total
phosphorus
(mg/l)
Total
nitrogen
(m g/l)
Suspended
solids (mg/l)
Hydrogen
carbonate
(mg/L)
Chloride
(mg/l)
Sulfate
(mg/l) pH
Electrical
Conductivity
(μs/cm)
Influent
water
2011-01-14 2.1 3.2 0.031 0.012 9.32 0.106 9.84 19.3 301 19.2 26.0 7.31 636
WME 2011-01-14 2.2 2.7 0.026 0.012 9.41 0.089 9.85 22.0 302 18.4 25.6 7.35 634
Influent
water
2011-02-23 2.1 2.1 0.008 0.010 13.79 0.015 13.70 0 325 23.7 39.8 7.64 770
WME 2011-02-23 2.4 2.4 0.007 0.020 13.94 0.020 14.17 0 331 23.8 39.4 7.63 770
Influent
water
2011-05-09 0.4 0.5 0.011 0.011 13.90 0.027 13.95 1 335 27.7 45.8 7.53 783
WME 2011-05-09 0.4 0.9 0.011 0.011 13.90 0.024 14.18 1 336 27.4 44.2 7.46 781
Influent
water
2012-01-11 1.0 1.3 0.007 0.017 10.85 0.021 11.14 0.0 356 24.4 35.7 7.67 780
WME 2012-01-11 0.9 1.1 0.006 0.021 10.92 0.018 11.33 0.5 347 23.8 35.1 7.56 772
Lab limit of
quantification (mg/l)
0.3 0.3 0.014 0.029 0.10 0.070 1.5
SRE 0.07 0.23 0.13 0.17 0.01 0.18 0.02 0.03 0.01 0.02 0.02 0.01 0.00
Lab analytical error 0.30 0.01 0.05
1031 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
collected at 10 separate times throughout the year and had a
coefficient of variation of 0.01. The error in the WME esti-
mation of discharge compared to the manual
measurements of discharge during the previously mentioned
10 measurement times throughout the year was 0.05. The
error was estimated using Equation (2), but replacing con-
centrations with discharges. The flows assessed ranged
from 0.04 to 0.5 l/s. Beyond approximately 0.7 l/s the plun-
ger remains opened continuously and therefore cannot be
used to estimate discharge. If there is a need to measure dis-
charge with the WME, then a slight magnetic force between
the plunger and the outflow pipe may be required to ensure
a proper closing.
DESIGN LIMITATIONS
During the operation of the WME, we found two functional
limitations and issues. The first occurred at a site with
very low discharge (<0.04 l/s) combined with a heavy
fine sediment load (up to 100 mg/l of silt). In this situation,
the WME did not flush frequently enough to continuously
remove all of the sediment transported into the WME. The
maximum sediment accumulation that we found was ∼4cm
of sediment deposited over a 2-week period.
The second issue was the accumulation of precipitated
calcium carbonate within the WME. Normal agricultural
practice in Austria (and many parts of the world) includes
periodic spreading of lime on the agricultural fields to
ensure that the soil does not become acidic. As with most
chemicals that are spread onto agricultural fields, the cal-
cium eventually makes its way through the various
pathways and into the main water bodies. During our first
8 months of operation, we had only minimal accumulation
of calcium carbonate, which caused minimal functional
issues for the WME. The first major rain storm after the
farmers applied the calcium bicarbonate caused significant
accumulations of calcium carbonate at nearly all of the
WME stations. The primary issue with the build-up of cal-
cium carbonate is that the flushing mechanism of the
plunger within the center shaft seizes from the shrinking
diameter in the center shaft and the additional friction.
The seizing occurred on approximately a third of the
devices, and the others had some degree of problems open-
ing properly from an incomplete plunger seal.
Both issues of the sediment accumulation at very low
flows and calcium carbonate cementation are situational
issues and should be dealt with as such. Periodic cleaning
of the WME is necessary to ensure optimal performance,
and the frequency of cleanings will be unique for each
installed WME due to the local environmental conditions.
During the period when calcium carbonate was accumulat-
ing rapidly, our WMEs were cleaned once every 2 weeks.
During the remainder of the year, the WMEs would only
need to be cleaned once every couple of months or in
some cases not at all.
CONCLUSIONS
The WME has been shown to provide a simple, flexible, and
effective housing for many types of water measurement
devices. The WME can function in many environments
that may be remote and without electricity for long periods
of time at nearly any range of flow. The WME ensures a
minimum internal water level and this ensures that the
internal monitoring equipment remains submerged even
when there is no flow into the enclosure. The limited diam-
eter of the inflow pipe and water volume in the WME
buffers the flow velocity from dramatic changes. The
device ensures that the sediment entering the enclosure
from the inflow will be conveyed through the enclosure
with minimal sediment accumulation. The device is pow-
ered purely from natural hydraulic forces, so it requires no
power source, and requires little additional maintenance
beyond periodic cleaning.
The functional assessments have shown that the WME
has a minimal effect on the chemistry of the water, and
with the addition of a small magnet the WME can also
measure discharge accurately up to 0.5 l/s. In certain
environments, the WME must be cleaned regularly to
ensure the proper functioning of the few moving parts
within the WME. We cannot emphasize enough the impor-
tance of regular maintenance for any water monitoring
device regardless of how simple and basic the device may be.
ACKNOWLEDGEMENTS
We would like to thank the Austrian Science Foundation for
funding our work as part of the Vienna Doctoral
Programme on Water Resource Systems (DK Plus W1219-
N22). We would like to thank Jan Podrouzek for his
frequent field assistance and tireless efforts installing the
WMEs. We would also like to thank Peter Haas and Ernis
Saracevic for their assistance and input during the labora-
tory testing of the WME.
1032 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
REFERENCES
Burns, D. A., McDonnell, J. J., Hooper, R. P., Peters, N. E., Freer,
J. E., Kendall, C. & Beven, K. Quantifying contributions
to storm runoff through end-member mixing analysis and
hydrologic measurements at the Panola Mountain Research
Watershed (Georgia, USA).Hydrological Processes 15 (10),
1903–1924.
Carrasco, M. E., Bautista, J. A. G. & Mateo, J. V. G.
Automated sequential monitoring of ammonium, phosphate
and nitrite in wastewater by multi-commutated peristaltic
and solenoid pumped flow system –a comparative study.
Chemia Analityczna 52 (5), 757–770.
Committee EH/3/2 Water analysis. Guidelines for the
determination of total organic carbon (TOC) and dissolved
organic carbon (DOC). BS EN 1484:1997.
Ensign, S. H. & Paerl, H. W. Development of an unattended
estuarine nutrient monitoring program using ferries as data-
collection platforms.Limnology and Oceanography: Methods
4, 399–405.
Harmel, R. D., Cooper, R. J., Slade, R. M., Haney, R. L. & Arnold,
J. G. Cumulative uncertainty in measured streamflow
and water quality data for small watersheds. Transactions of
the ASABE 49 (3), 689–701.
Kaelin, D., Rieger, L., Eugster, J., Rottermann, K., Bänninger, C.
& Siegrist, H. Potential of in-situ sensors with ion-
selective electrodes for aeration control at wastewater
treatment plants.Water Science and Technology 58 (3),
629–637.
Kestel, S., Gray, M. & Lee, G. Effect of ionic strength on ion
selective electrodes in the activated sludge process.
Proceedings of the Water Environment Federation WEFTEC
2010: Session 91 through Session 100, 6955–6960.
Kim, H.-J., Sudduth, K. A. & Hummel, J. W. Soil
macronutrient sensing for precision agriculture.Journal of
Environmental Monitoring 11 (10), 1810–1824.
Legnerová, Z., Solich, P., Sklenár
ová, H., Šatínský, D. & Karlícek,
R. Automated simultaneous monitoring of nitrate and
nitrite in surface water by sequential injection analysis.Water
Research 36 (11), 2777–2783.
Newman, I. A. Ion transport in roots: measurement of fluxes
using ion-selective microelectrodes to characterize
transporter function.Plant, Cell & Environment 24 (1), 1–14.
Roser, D., Skinner, J., LeMaitre, C., Marshall, L., Baldwin, J.,
Billington, K., Kotz, S., Clarkson, K. & Ashbolt, N.
Automated event sampling for microbiological and related
analytes in remote sites: a comprehensive system. Water
Science and Technology: Water Supply 2(3), 123–130.
Stenholm, Å., Eriksson, E., Lind, O. & Wigilius, B.
Comparison of continuous flow analysis including
photometric detection and ion-selective electrode
potentiometry for the measurement of ammonium nitrogen
in wastewater.International Journal of Environmental
Analytical Chemistry 88 (3), 165–176.
Tan, C. S., Drury, C. F., Soultani, M., Van Wesenbeeck, I. J., Ng,
H. Y. F., Gaynor, J. D. & Welacky, T. W. Effect of
controlled drainage and tillage on soil structure and tile
drainage nitrate loss at the field scale.Water Science and
Technology 38 (4-5-5 pt 4), 103–110.
First received 16 March 2012; accepted in revised form 15 October 2012
1033 M. Exner-Kittridge et al.|A simple and flexible device for housing water monitoring sensors Water Science & Technology |67.5 |2013
