United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S7-85/011 May 1985
c/EPA Project Summary
Porous Dike Intake Evaluation
Barry A. Ketschke and Richard C. Toner
A small-scale porous dike test facility
was constructed and continuously
operated for 2 years under field condi-
tions. Two stone dikes of gabion con-
struction were tested: one consisted of
7.5 cm stones; and the other, 20 cm
stones. Approach velocity was set at 3
cm/sec. Using a test flume, laboratory
studies were also conducted on the
avoidance response of fish to a porous
dike intake.
Flow through a porous dike, induced
by a hydraulic head of about 61 cm,
depended on the cross-sectional area
(tidal height). Throughout the experi-
ment, flow resistance changed, depend-
ing on the fouling. Increases in flow
resistance resulted in lower flow rates
at constant hydraulic head. Flow rates
could be increased by backflushing the
stone dike. Throughout the experiment,
the porous dike showed a steady ac-
cumulation of silt and organic matter,
but the flow rates tended not to decrease
with time.
Analysis of field and laboratory data
showed that the number of zooplankton
is reduced passing through a porous
dike by a combination of physical filtra-
tion and cropping. No zooplankton
avoidance was substantiated. Analysis
also indicated that the number of pas-
sively drifting larval fish is reduced
passing through a porous dike by filtra-
tion or cropping. No conclusive field
evidence was obtained to explain if
actively swimming larval fish avoid
passing through the dike; however,
limited laboratory results suggest avoid-
ance may occur. Results with juvenile
and adult fish show that they avoid
passing through a porous dike.
This Project Summary was developed
by EPA's Air and Energy Engineering
Research Laboratory, Research Triangle
Park. NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).
Introduction
Several industries, particularly the
steam-electric industry, have large re-
quirements for cooling water. All cooling
water systems, whether they recirculate
the water (as in cooling towers) or not (as
in once-through cooling systems), must
have an intake system. These intake
systems direct water, including the
aquatic organisms contained in the water,
into the industrial facility. The combina-
tion of screening, heat, mechanical action
and, in some cases, chemicals added to
the cooling water, can result in mortality
to aquatic organisms.
The Federal Clean Water Act Amend-
ments of 1972 ultimately require that all
cooling water intakes be equipped with
the best technology available to minimize
the environmental impacts on aquatic
organisms. After the Clean Water Act
became law, the USEPA and the industrial
users of cooling water began to examine
and experiment with cooling water intake
designs.
One method considered feasible was a
rock (or porous) dike. Conceptually, the
porous dike would be similar to a stone
breakwater and surround the intake. All
cooling water would then flow through
the rock dike. By selecting a very low
approach velocity, about 3 cm/sec (0.1
ft/sec), and relatively small stones, 7.5
cm (3 in.) to 20 cm (8 in.), it was theorized
that the system could act as a physical
barrier for large mobile aquatic organisms
-------
(e.g., fish) and as a behavioral barrier for
small weak swimming aquatic forms
(e.g., larval fish and microscopic animal
plankton).
For a porous dike system to be con-
sidered a feasible intake, it was necessary
to establish that it was in fact a barrier to
aquatic organisms and that the system
would maintain porosity and not clog.
The research covered in this report was
designed to examine these two questions.
Specifically, the research was designed
to examine the ability of a porous dike to
keep zooplankton, larval fish, juvenile fish,
and adult fish out of industrial cooling
water systems. A second objective was to
examine the hydraulic characteristics and
clogging rates of a porous dike.
Materials and Methods
General Procedures
All hydraulics, fouling, and field screen-
ing performance testing was conducted
at a small-scale porous dike test facility,
at Brayton Point Station, Somerset MA.
The porous dike test facility (Figure 1)
was a reinforced concrete and steel
structure 6.4 m wide, 18.3 m long, and
6.1m deep. The chamber was a concrete
box, open at the top and front and divided
into three cells by wooden timbers. Each
cell was 1.8 m wide. A single axial flow
pump in the back (downstream) end of the
box drew water through the structure. At
the front (upstream) end of the box,
gabions were placed in the first and third
cells. One cell had gabions filled with 7.5
cm stone, and the other had gabions filled
with 20 cm stone. Each gabion was 1.8 m
wide, 0.9 m high, and 0.9 m deep. They
were placed on top of each other to form a
wall approximately 4.3 m high. In the 20
cm test section, three rows of gabions
made a 2.7 m thick section. In the 7.5 cm
test section, two rows made a 1.8 m thick
section. The center cell was sealed and
not used. Flow was regulated by flow-
control baffles. Approach velocities were
initially set at about 3 cm/sec(0.1 ft/sec).
The test facility went into full operation
on July 16,1979.
Hydraulic Performance
Hydraulic characteristics were meas-
ured by recording over the 2 years of
operation the frictional losses through
the stones or head loss (h) and the
volumetric flow (Q) through the stones.
Additionally, special hydraulic studies
were conducted to address the efficiency
of back-flushing and the relationship
between increased head and volumetric
flow.
r- a
Impoundment 4$^ •
Net 'f.
®
Flow^
20 cm Stones L,
7.5 cm Stones
\xC
\^j)
Flow^
1 Plan
. . ' • .'•••• •••" '. -'' '• •' .. / '*/:
lEC r '
/?
° •<&* '© 'i
**i &y /.sm ^
h
Flow
- Control
Gates
'
":":•
In
.
U-T7
'•;.
,-{v.-.|
/7
]
/Ax/a/ Pump
-18.3 m-
-6.4m-
K
°
*•*.-.
9 -V*
•*
®. ®. © -Biological Sampling
Stations
h-Head Staff
v - Velocity/Flow
6.1 m
~ Fouling Boxes
Section a-a
Figure 1. Porous dike test facility.
Head loss was measured by recording
the water level at positions upstream of
the gabions, downstream of the gabions,
downstream of the flow control baffle,
and downstream of the pump (Figure 1).
Additionally, some measurements of h
were made in the gabion slots.
Water level was measured with 4.6 m
(15 ft) measuring staffs which were fixed
to the walls of the test facility. In the
gabion slots, water level was measured
using a float and graduated rod. Head loss
through the stones was calculated as the
difference between the upstream water
level and the downstream water level for
each test station.
Volumetric flow was measured by
recording the velocity of water passing
under the known area of the flow control
baffle gate. Velocity under the flow control
baffle gate was measured in each test
section at six fixed locations using a
portable current meter. The current meter
was fixed to a 6.1 m (20 ft) pole equipped
with a mounting bracket which fit into a
slot on the baffle gate. This setup ensured
that measurements were made at the
same depth and location on each survey.
Two backflush studies were conducted—
in May 1981 and in July 1981. In both
studies, the methodology was the same.
At time zero, the flow volume and corre-
sponding head loss values were meas-
ured throughout a tidal cycle. These flows
and heads represented the initial or
baseline values. After initial flows were
measured, the upstream face of the 20
cm stone section was cleaned using a
diver-operated pump. The speed of the
gasoline-engine-powered pump was ad-
justed to yield a suction which penetrated
no more than 5 cm (2 in.) into the dike.
After this face cleaning, flow volumes
and head readings were taken for com-
parison to the initial Q and h readings.
This face cleaning was designed to
determine if the increased flow resistance
was due to a surface fouling film. After
the flow and head readings were taken,
following the face cleaning, backflushing
was initiated. Backflushing was achieved
by turning the axial flow pump off. The
facility was allowed to backflush for
varying periods, upto216 hours(cumula-
tive). After each backflush, the pump was
operated briefly, and the h and Q readings
recorded.
The relationship between increased
head differential and flow volume was
examined by augmenting the flow rate of
the axial flow pump with three gasoline-
engine-powered pumps. Pumps were
-------
added in sequence to increase the head
differential and the resultant flow meas-
ured.
Fouling Rate
The fouling rate was measured by the
change in weight and void space of the
gabions. The changes in gabion weight
and void space were measured with
fouling boxes, small removable sections
of the gabions. Fouling boxes were 22 cm
x 22 cm x 59 cm stainless steel boxes
open at each end and filled with 7.5 cm or
20 cm stones. These boxes were in
drawers between the bottom and next-to-
bottom gabions on the upstream and
downstream face of each test section at
start-up. Prior to installation, the weight
and displacement volume of each box
was measured and recorded. Every 3
months, one box from the upstream face
of each test section was removed and the
weight and displacement volume meas-
ured and recorded. After each fouling box
was removed, all biofouling was identified
and measured. All non-living detrital
material was collected and weighed.
Laboratory Screening
Performance Tests
Laboratory screening tests for zooplank-
ton, larval fish, and juvenile and adult fish
were conducted in a T-shaped flume
(Figure 2). Dimensions of the long arm of
the T were 3.0 m x 0.5 m x 0.5 m. The
short arm measured 2.4 m xO.5 m xO.5 m.
Valves, between a 10 hp (7.46 kW) pump
and the flume, controlled both current
velocity and direction of flow. Velocity
could be maintained between 0.5 and
about 30 cm/sec (0.02 to 1 ft/sec). The
flume was directly connected with fish
holding facilities and a mechanical and
biological filter system which permitted
test animals to acclimate in the same
water used in flume experiments. Light
level and water temperature were also
controlled.
Laboratory Zooplankton Screening
Tests. Zooplankton tests were run in the
flume with valve openings set to yield an
equal volume of water flowing down each
arm toward locations B and C (Figure 2).
Under this arrangement, water flowed
down the short arm (from location A): at
the intersection, half the water flowed
down the long arm; and the other half,
down the short arm.
Live zooplankton of known species
composition and number were placed in
the flume at location A. Under test
conditions, a small rock dike 51 cm thick
made of 20 cm stones was placed in the
Removable Gabion
A, B, C - Sampling
Locations
a, b- Flow Readings
Figure 2. Laboratory test flume.
long arm at the intersection. Under
control conditions, the stones were
removed.
The zooplankton introduced into the
short arm at location A were captured
downstream of the intersection in both
the long (C) and short (B) arms. Current
velocity was maintained at 3 cm/sec.
All preserved samples were analyzed
by microscope as to type and number of
species present. All density data were
converted to the ratio of the number
captured over the number introduced
before gabion and after gabion. These
ratios were used in a paired t-test to
determine whether any differences exist
between the control and experimental
conditions.
Laboratory Fish Screening Tests. Ex-
periments designed to assess the porous
dike concept were conducted with the T-
shaped flume (Figure 2). A small rock
dike, 51 cm thick and made of 20 cm
stones, was constructed where the long
arm of the flume joined the short arm. The
rocks were carefully packed between
stainless steel racks or gabion so that
channels through which water could flow
unimpeded did not exist along the sides or
bottom of the flume. Rock panels affixed
to the walls alongside the gabion gave the
appearance of a continuous rock wall
along the back of the forward chamber. In
certain experiments, rocks, similar to
those in the dike, were widely scattered
on the bottom. In others, 10 to 15 mm
gravel was spread in a single layer on the
bottom. Water current was directed from
A to B (Figure 2) across the face of the
rock. Approximately half the volume
passed in front of the stone dike, and the
other half passed through the dike.
To ensure that the behavior of test
animals was not influenced by an ob-
server, a viewing screen was constructed
around the face of the short arm.
Current velocity was generally main-
tained at 2, 3, or 6 cm/sec (0.07, 0.1 or
0.2 ft/sec) measured at 1 m downstream
of the rockdike. Velocities under 3 cm/sec
were measured by timing the movement
of dye or a neutrally buoyant, water-filled
table tennis ball since both flowmeters
were insensitive to such low velocities.
Light was generally maintained at 65
lux provided by 4 incandescent lights
mounted above the small arm of the
flume. Two additional incandescent lights
provided about 45 lux above the long arm.
A rheostat was used to reduce light levels;
for 24-hour studies, a timer incorporating
a twilight period was used.
-------
Since water temperature in the flume
continually increased, due to friction
between the water and flume surfaces as
well as pump work, temperatures were
maintained within 2.5°C of the initial
acclimation temperature by exchange
with cooler, ambient seawater or ex-
change between the flume and a separate
insulated reservoir containing two cooling
units.
The basic experimental procedure con-
sisted of introducing small fish or larvae
into the short arm of the T at location A
and recording their behavior for periods
of from 30 minutes to several hours.
Before release, each sample of fish was
acclimated in the current for 20-30
minutes in either a screened 8-liter pail or
screened 1 -liter beaker with removable
bottom.
An additional study was conducted to
observe the filtering capacity of the stone
in the porous dike test facility. A fouling
box containing 20 cm stone was removed
from the test facility at Brayton Point,
transferred to the flume, and placed at the
intersection. The leading edge of the box
was fitted into a wooden panel cut to fit
the exact dimensions of the box mouth so
that all water flowing down the long arm
of the flume passed through the box. A
small 0.33 mm mesh net with cod end
was tied to the back edge of the box to
catch all material passing through it. Flow
rate through the box was set at 3 cm/sec.
Samples of preserved larval winter
flounder and labrid eggs were released at
the face of the box while the flume was
operating. Flow was maintained for 7
minutes at which time the flume pump
was shut off, the net removed and washed
down, and the sample preserved until
counted.
Field Screening
Performance Tests
All field screening tests for zooplankton
and larval, juvenile, and adult fish were
conducted at the porous dike field test
facility (Figure 1).
Field Zooplankton Screening Tests.
Zooplankton field studies were conducted
on 90 days between July 1979 and July
1981.
Evaluation of the effectiveness of a dike
in reducing zooplankton entrainmentwas
based on comparisons of the population
density at locations upstream and down-
stream of, and within the dike. Sequential
sampling was carried out during daylight
at locations A, B, and C for both the 7.5
and 20 cm rock channels (7.5 A, 7.5 B, 7.5
C, 20 A, 20 B, and 20 C—see Figure 1)
eight times each sampling day. Collec-
tions at each pair of locations (7.5 A—20
A, etc.) were made simultaneously using
identical gear for both channels. Each
sample was taken by pumping 500 liters
of water through No. 30 (59 (im) mesh
nets using gasoline-engine-powered
pumps. To ensure that all of the water
column was sampled, one-third of each
sample was pumped from the bottom
third of the water column, from mid-
depth, and from the top third of the water
column, and eight samples from each
sampling location in each channel were
combined into a single composite sample
which represented the zooplankton popu-
lation for that location. During the first
year, sampling was conducted weekly,
with 48-hour die! sampling once a month.
During the second year, sampling was
reduced to biweekly, with a 24-hour diet
sampling once a month.
In the laboratory, each sample was
reduced in volume to a concentration of
300-500 organisms/ml, and three 1 ml
aliquots were removed and analyzed as to
type and number of each organism pre-
sent. To facilitate analysis, volumes were
kept under 1,000 ml. If necessary, sam-
ples were split with a plankton splitter.
Data were analyzed for density differ-
ences.
Field Larval Fish Screening Tests.
Ichthyoplankton densities on the upstream
and downstream sides of the porous dike
were sampled between August 1979 and
July 1981 during March, April, May, June,
July and August—when ichthyoplankton
were most abundant.
Ichthyoplankton samples were taken
with a gas-driven trash pump using 10 cm
intake and discharge hoses and a 0.33 mm
mesh plankton net. To improve the condi-
tion of ichthyoplankton, the last 2.4 m of
discharge hose was increased to 15 cm
diameter and the net was submerged.
Pumping duration per sample ranged from
15 to 30 minutes depending on the
abundance of ichthyoplankton in the
samples. During each pumping period,
the intake hose was moved at set intervals
so that surface, mid-depth, and near-
bottom were sampled equally.
Pumping volume was determined by
recording the time necessary to fill a 757-
liter box suspended in the same position
as the sampling net. A tachometer,
accurate to within ±2%, was used with
the pump to ensure operational accuracy
between sampling and calibration.
In 1979 and 1980, sampling was
conducted 2 days per week, 1 day in the
7.5 cm channel, and 1 day in the 20 cm
channel. In July and August 1980, sam-
pling frequency was increased to 3 days
per week to compensate to some extent \
for time lost when the facility was shut
down for repair. In 1981, sampling was
again conducted 2 days per week with 1
(March—mid-June) or 2 (mid-June—July)
night sampling periods, in addition to the
day periods.
On each sampling day, samples were
taken at locations A and C (Figure 1).
Generally 4 to 10 sets of samples were
taken each day, depending on pumping
duration. Although the original sampling
design included collecting within the slots
between gabions (B positions in Figure 1),
a problem developed when the weight of
stone caused the gabions to bulge into
the slots. In the 20 cm stone channel, the
slots were so narrow that the 10 cm hose
could not be lowered into the water.
Initially, sampling was possible at position
B of the 7.5 channel, but only near high
tide because the hose could not always be
fitted far enough into the slot to reach
water level. Beginning in May 1980, an
additional intake hose was obtained for
the B position of the 7.5 cm channel. This
hose was worked into the slot and left in
place so that the samples could be taken
regularly at that position. Because the
hose could not be raised or lowered in the
slot, sampling was restricted to one
depth—about 1.2 m above the bottom.
All samples were preserved in 10%
Formalin and returned to the laboratory
for microscopic analysis. Fish eggs and
larvae were identified to the lowest
distinguishable taxonomic category and
counted.
Data were analyzed for density differ-
ences.
FieldFish Screening Tests. Field finfish
experiments were conducted at the
porous dike between September 1979
and July 1981, using the impoundment
enclosures shown in Figure 1. The en-
closures were constructed at 4.7 mm
nylon mesh including a complete bottom,
but were open where they joined the dike.
Each enclosure was 4.3 m high. The
mouth of each enclosure was held open
by a metal frame which fit tightly within
U-shaped slots in front of each channel.
The offshore section of each net was held
in place by a system of haul lines and
anchor poles. The head ropes were fitted
with floats and were pulled tight to keep
them above high water.
Fish to be tested were collected by
beach seine, otter trawl, lift net, or on the
revolving screens at Brayton Point Power
Station. Typically, the nets were set, a
known number of test fish were released
in each enclosure, and the system was
-------
left undisturbed for a period of time
ranging generally from 24 to 72 hours. At
the end of the test period, the nets were
hauled, mouth first, and the fish remain-
ing inside were counted and measured.
Fish were selected by size such that
they were too large to escape through the
enclosure meshes and yet small enough
to enter the spaces between rocks of the
dike if they chose to do so. During the
winter months, tests were run whenever
fish were available and weather permit-
ted.
Since early in the study, it was not
possible to determine whether missing
fish had escaped or entered the stones.
Therefore, a method of seining the down-
stream section of each channel was
developed in July and August 1980. All
fish introducedto the impoundments from
that point on were finclipped so they
could be easily identified if subsequently
collected in the downstream channels.
The seine measured 3 m x 3 m, was
constructed of 4.7 mm delta mesh, and
was lashed along the vertical edges to
two 3 m long wooden poles. The bottom
edge of the seine was weighted with
chain; the bottom of each pole was
weighted with lead. To assist in hauling
the seine over the length of the channel,
haul lines were fixed to the top and
bottom of each pole. To confine fish in the
channels during seining, 4.7 mm mesh
screens were constructed to slide down
in slots fixed to the channel walls in front
of the openings under the baffle gates.
Generally, three seine passes were made
whenever an impoundment test was
conducted.
Conclusions
The results of studies of the hydraulic
characteristics of flow through a rock dike
demonstrated that a low flow velocity of 3
cm/sec can be induced and maintained
by a relatively low hydraulic head of
approximately 61 cm. Increasing hydraulic
head and cross-sectional area result in
increased volumetric flow through the
stones. In tidal water, the decrease in
flow due to decreasing depth is greater
than would be predicted based on the
reduction cross-sectional area, indicating
higher flow resistance in the lower stone
sections.
Flow volume induced by a relatively
constant head varied throughout the
experiment. Some seasonal trends were
evident with lowest flows occurring dur-
ing the warmest months. Flow volume
increased both seasonally and in response
to backflushing. Flow volume could be
maintained by backflushing.
The major resistance to flow was in the
upstream 0.9 m of dike. Organic matter
and debris continued to accumulate with-
in the stones through the 2-year experi-
ment. About 70% of the fouling material
was detritus, including organic debris.
Variations in flow appeared to be co-
related more to seasonal surface fouling,
particularly hydroids and shelled forms,
than to the accumulation of material
within the stones. About 5 yd3 (0.382 m3)
of silt and debris accumulated downsteam
of the dike over the 2-year test period.
There was no difference in the hydraulic
performance between the two stone sizes
tested (7.5 cm and 20 cm). The fouling
organisms in the two stone dikes showed
some differences: the 20 cm stone diking
had more larger shelled organisms, and
the 7.5 cm dike had more small attached
organisms.
The number of zooplankton per volume
of water was reduced by passing through
a porous dike. The amount of reduction
was less during the colder months. The
reduction was attributed to physical filtra-
tion and predation by the fouling com-
munity. No evidence, either in the field or
the laboratory, could substantiate a zoo-
plankton avoidance response.
The numbers of fish eggs and larvae
per volume of water were reduced, in
most cases, passing through the porous
dike. The reduction in the density of
passively drifting fish eggs was attributed
to physical filtration and perhaps preda-
tion by the fouling organisms within the
stones. The reduction in the number of
passively drifting larvae such as winter
f lou nder (Pseudopleuronectes americanus)
was attributed to physical filtration and
perhaps cropping. Winter flounder were
shown to exhibit no directional swimming
ability or avoidance response in laboratory
test flume experiments; the reduction in
density observed in the field was, there-
fore, attributed to filtration or predation.
No conclusive evidence was obtained in
the field to establish whether or not
density differences for actively swimming
larvae were due to avoidance or filtration/
predation. Laboratory tests did show that
some species of larvae exhibit a direc-
tional swimming ability and avoidance
response.
Results of both laboratory and field
tests have demonstrated that the porous
dike is an effective barrier to juvenile and
adult fish.
Recommendations
The role of filtration, predation, and
avoidance in reducing the density of larval
fish needs to be investigated further. This
investigation could include an evaluation
of the effect of flow path length on filtra-
tion capacity.
A larger scale test of a porous dike
intake is also needed, to address the
specific hydraulic design parameter
needed for widespread commercial appli-
cation. These studies could examine
various structural designs that would
capitalize on the localized fouling and
backflushing potential.
. S. GOVERNMENT PRINTING OFFICE: 1985/559 111/10851
-------
B. Ketschkeis with New England Electric System, Westboro, MA;R. Toner is with
Marine Research, Inc.. Fa/mouth, MA.
Julian W. Jones is the EPA Project Officer (see below).
The complete report, entitled "Porous Dike Intake Evaluation," (Order No. PB
85-185 908/AS; Cost: $14.50, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
OCOC329 PS
U S ENVIR PROTECTION AGENCY
REGION 5 LIBRARY
230 S DEARBORN STREET
CHICAGO IL £0604
------- |