MEMORANDUM

Date

October 17, 2008

To

Marla Smith, Work Assignment Manager

From

Bob Lordo, Battelle Work Assignment Leader

Chris Sroka, Battelle Statistician

Subject

EPA Contract No. EP-C-05-030, Work Assignment No. 3-13:  

316b: Laboratory Test Data Related to Entrainment 

This memorandum presents entrainment data from studies that assessed the
performance of fine mesh screens in reducing impingement mortality and
entrainment under controlled laboratory conditions.  These entrainment
data originated from documents which we have been reviewing on this work
assignment and its predecessor in support of the Section 316(b)
rulemaking effort.  

This memorandum also summarizes the results of statistical analyses to
evaluate whether slot width and slot velocity have a significant impact
on the percentage of organisms that are entrained through the screens.  

Measure of Performance Considered in the Laboratory Studies

Ideally, we would use the laboratory study data to derive the same
performance measure as that used in the field studies:  percentage
reduction in entrainment.  Recall that this measure requires
measurements from paired density samples taken in front of and behind a
screened intake.  However, none of the encountered laboratory studies
performed this type of sampling.  Instead, the laboratory studies
measured the extent of entrainment (and impingement mortality)
associated with the installed screens within the test flume.  Therefore,
the percentage of organisms that were entrained through the screen
(i.e., percent entrainment) is the performance measure used in the
summaries and analyses of laboratory data in this memorandum.  Note that
because field and laboratory studies are utilizing different entrainment
performance measures, it is not possible to make direct comparisons in
the values of the performance measures between field and laboratory
studies. 

Sources of Entrainment Data from Laboratory Studies

The summarized entrainment data used in this evaluation are documented
in Tables 1 through 3.  Table 1 summarizes entrainment for all organisms
(i.e., all life stages), while Table 2 considers entrainment data for
eggs only, and Table 3 documents entrainment for larvae (non-eggs) only.
 These data represent summaries calculated across species of the data
presented within the source specified in the first column.  The
summaries are presented by slot size and, when specified, slot velocity
and screen orientation (parallel or perpendicular to water flow).  The
Study ID is assigned to the given data within Battelle’s performance
evaluation database.  (The data from EPRI [2003] currently exist in
separate Excel worksheets, as they were captured in electronic form from
a PDF version of the study report.)Table 1.	Percent Entrainment of All
Organism Types by Screen Orientation, Slot Size, and Slot Velocity, as
Reported from Controlled Laboratory Studies

Data Source	Study ID	# Species	Screen Orientation*	Slot Size (mm)	Slot
Velocity (m/s)	# Entrained**	# Released	Avg. % Entrained**	Avg. %
Retained

Appendix C of EPRI, 2003	2001 WW lab study	1	0	0.50	0.15	39.70	2,000
1.99	 

	2001 WW lab study	1	0	0.50	0.30	249.50	1,900	13.13	 

	2001 WW lab study	1	0	1.00	0.15	255.20	1,800	14.18	 

	2001 WW lab study	1	0	1.00	0.30	357.60	1,800	19.87	 

	2001 WW lab study	1	0	2.00	0.15	504.60	1,800	28.03	 

	2001 WW lab study	1	90	0.50	0.15	2.20	1,800	0.12	 

	2001 WW lab study	1	90	0.50	0.30	2.20	1,800	0.12	 

	2001 WW lab study	1	90	1.00	0.30	304.10	1,050	28.96	 

	2001 WW lab study	1	90	2.00	0.15	893.70	1,800	49.65	 

Appendix D of EPRI, 2003	2002 WW lab study	4	 NS	0.50	0.15	649.60	5,688
11.42	 

	2002 WW lab study	5	 NS	0.50	0.30	471.70	4,285	11.01	 

	2002 WW lab study	3	 NS	1.00	0.15	1477.40	2,525	58.51	 

	2002 WW lab study	3	 NS	1.00	0.30	1859.70	2,773	67.06	 

	2002 WW lab study	1	 NS	2.00	0.15	900.40	1,125	80.04	 

	2002 WW lab study	2	 NS	2.00	0.30	427.60	1,874	22.82	 

Table 2-4 of EPRI, 2008	116	1	90	0.50	 NS	9.00	120	7.50	92.50

	116	4	90	0.50	 NS	57.89	1,250	4.63	95.37

	116	5	90	0.50	 NS	36.54	1,695	2.16	97.84

Table 1 of Tomljanovich et al., 1978	248	5	90	0.50	 NS	 	 	0.36	99.64

	249	6	90	1.00	 NS	 	 	19.97	80.03

	250	9	90	1.30	 NS	 	 	35.67	64.33

	251	4	90	1.80	 NS	 	 	20.20	79.80

	252	4	90	2.50	 NS	 	 	32.60	67.40

Page 401 of Hanson et al., 1977	253	1	 NS	1.01	 NS	16.00	100	16.00	 

* 0=parallel to flow; 90=perpendicular to flow.

** Adjusted for collection efficiency (CE) for WW lab studies.

NS = Not specified

Table 2.	Percent Entrainment of Eggs by Screen Orientation, Slot Size,
and Slot Velocity, as Reported from Controlled Laboratory Studies

Data Source	Study ID	# Species	Screen Orientation	Slot Size (mm)	Slot
Velocity (m/s)	# Entrained*	# Released	Avg. % Entrained*

Appendix C of EPRI, 2003	2001 WW lab study	1	0	0.50	0.15	0.00	900	0.00

	2001 WW lab study	1	0	0.50	0.30	0.00	900	0.00

	2001 WW lab study	1	0	1.00	0.15	0.00	900	0.00

	2001 WW lab study	1	0	1.00	0.30	0.00	900	0.00

	2001 WW lab study	1	0	2.00	0.15	0.00	900	0.00

	2001 WW lab study	1	90	0.50	0.15	0.00	900	0.00

	2001 WW lab study	1	90	0.50	0.30	0.00	900	0.00

	2001 WW lab study	1	90	1.00	0.30	0.00	600	0.00

	2001 WW lab study	1	90	2.00	0.15	21.00	900	2.33

Appendix D of EPRI, 2003	2002 WW lab study	1	 NS	0.50	0.15	0.00	1,122
0.00

	2002 WW lab study	2	 NS	0.50	0.30	100.20	1,737	5.77

	2002 WW lab study	1	 NS	2.00	0.30	407.60	1,125	36.23



* Adjusted for collection efficiency (CE)

NS = Not specified

Table 3.	Percent Entrainment of Larvae (non-eggs) by Screen
Orientation, Slot Size, and Slot Velocity, as Reported from Controlled
Laboratory Studies

Data Source	Study ID	# Species	Screen Orientation*	Slot Size (mm)	Slot
Velocity (m/s)	# Entrained**	# Released	Avg, % Entrained**	Avg. %
Retained

Appendix C of EPRI, 2003	2001 WW lab study	1	0	0.50	0.15	39.70	1,100
3.61	 

	2001 WW lab study	1	0	0.50	0.30	249.50	1,000	24.95	 

	2001 WW lab study	1	0	1.00	0.15	255.20	900	28.36	 

	2001 WW lab study	1	0	1.00	0.30	357.60	900	39.73	 

	2001 WW lab study	1	0	2.00	0.15	504.60	900	56.07	 

	2001 WW lab study	1	90	0.50	0.15	2.20	900	0.24	 

	2001 WW lab study	1	90	0.50	0.30	2.20	900	0.24	 

	2001 WW lab study	1	90	1.00	0.30	304.10	450	67.58	 

	2001 WW lab study	1	90	2.00	0.15	872.70	900	96.97	 

Appendix D of EPRI, 2003	2002 WW lab study	4	 NS	0.50	0.15	649.60	4,566
14.23	 

	2002 WW lab study	3	 NS	0.50	0.30	371.50	2,548	14.58	 

	2002 WW lab study	3	 NS	1.00	0.15	1477.40	2,525	58.51	 

	2002 WW lab study	3	 NS	1.00	0.30	1859.70	2,773	67.06	 

	2002 WW lab study	1	 NS	2.00	0.15	900.40	1,125	80.04	 

	2002 WW lab study	1	 NS	2.00	0.30	20.00	749	2.67	 

Table 2-4 of EPRI, 2008	116	1	90	0.50	 NS	9.00	120	7.50	92.50

	116	4	90	0.50	 NS	57.89	1,250	4.63	95.37

	116	5	90	0.50	 NS	36.54	1,695	2.16	97.84

Table 1 of Tomljanovich et al., 1978	248	5	90	0.50	 NS	 	 	0.36	99.64

	249	6	90	1.00	 NS	 	 	19.97	80.03

	250	9	90	1.30	 NS	 	 	35.67	64.33

	251	4	90	1.80	 NS	 	 	20.20	79.80

	252	4	90	2.50	 NS	 	 	32.60	67.40

Page 401 of Hanson et al., 1977	253	1	 NS	1.01	 NS	16.00	100	16.00	 

* 0=parallel to flow; 90=perpendicular to flow.

** Adjusted for collection efficiency (CE) for WW lab studies.

NS = Not specifiedThe data in Tables 1 through 3 originate from the
following four documents:

EPRI (2003) is a biological evaluation of cylindrical wedgewire screens
performed at Alden Research Laboratory’s Fish Testing Facility.  Two
distinct sets of tests were performed that involved releasing a
specified number of larvae/eggs into a test flume.  One test was
conducted in 2001, with striped bass larvae (length range:  3.5 to 10.6
mm; width range:  0.1 to 2.0 mm) and a surrogate egg type (diameter
range: 3.6 to 6.3 mm), whose objective was to determine if the test
facility and procedures could be used to evaluate effectiveness of the
various screen types.  After necessary modifications were made to the
procedures based on 2001 testing, more comprehensive testing was done in
2002 that considered additional species that exist within a wide range
of water body types and that represent a greater range of swimming
capabilities.  Eggs used in the 2002 study represented alewife (diameter
range:  0.5 to 0.9 mm) and white sucker (diameter range:  2.8 to 3.5
mm), while larvae represented bluegill, carp, rainbow smelt, winter
flounder, white sucker, and yellow perch.  Mean lengths of these larvae
ranged from 6.1 mm (winter flounder) to 18.5 mm (bluegill), while mean
widths ranged from 0.8 mm (yellow perch) to 5.4 mm (bluegill).  Both of
the 2001 and 2002 studies controlled for the following:

Slot size (0.5, 1.0, 2.0 mm)

Slot velocity (0.15, 0.30 m/s)

Channel (or approach) velocity (0.08, 0.15, 0.30 m/s).

(Not all species of larvae were tested at all slot size/velocity
combinations in the 2002 study.)  In addition, the 2001 study collected
data for two different screen orientations:  parallel (radial slots) and
perpendicular (axial slots) to the water flow.  In 2001, three test
replicates were performed for each set of test conditions, while up to
five replicates were performed in 2002 testing.  The test flume channel
in 2001 testing was 1.5 m wide, and water depth was 1.3 m.  In 2002
testing, flume width and water depth were both 1.8 m.

The 2001 and 2002 data are summarized separately.  In both tests,
entrainment data were reported after being adjusted for net collection
efficiency (as determined from releasing known amounts of fish or eggs
directly into the entrainment collection tank).

EPRI (2008) is a two-year biological evaluation of fine mesh traveling
screens whose main objective was to determine retention and
post-impingement survival of larval fish on the screens.  These tests
were performed in a closed-loop testing facility at Alden Research
Laboratory.  The test flumes were 30.5 cm wide and contained water 30.5
cm deep, and fine-mesh screens were mounted perpendicular to the flow. 
Fish larvae were introduced to the system, then the study evaluated the
length of time at which the larvae could be impinged on the screens
before being rinsed off the screens and bypassed to holding tanks.  The
data presented in this memorandum originate from an investigation of the
ability of 0.5 mm mesh screens to prevent the entrainment of larval fish
that occurred within impingement survival testing in Phase I (in 2006). 
When using 0.5 mm mesh screens, survival testing involved performing
four replicates for selected combinations of the following:  

Larval species (blue catfish, channel catfish, emerald shiner, fathead
minnow, and smallmouth bass)

Approach velocity (0.152, 0.35, and 0.61 m/s)

Duration of impingement on the screens (4, 8, 16 min.).

The lengths of the larvae ranged from 6.9 to 23.4 mm (Table 2-3 of EPRI,
2008).

The study in Tomljanovich et al. (1978) was performed in a fine-mesh
screening test flume (2.4 m wide x 1.2 m deep) at the Tennessee Valley
Authority (TVA) to evaluate the extent to which post-impingement
survival of fish larvae were related to impingement duration, fish
species and size, approach water velocity, and screen mesh size.  The
study measured percent retainment of the larvae within various numbers
of tests for selected combinations of the following:

Species (jewelfish cichlid, threadfin shad, golden shiner/fathead
minnow, white sucker, channel catfish, striped bass, bluegill,
smallmouth bass, largemouth bass, walleye)

Mesh size (0.5, 1.0, 1.3, 1.8, 2.5 mm) of screens oriented perpendicular
to flow.

According to Table 1 of Tomljanovich et al. (1978), approach velocities
ranged from 0.152 to 0.46 m/s among the tests, and test duration ranged
from 0.5 to 16 minutes.  Mean larval lengths ranged from 6.2 (striped
bass) to 39.5 mm (threadfin shad), while mean widths ranged from 0.7
(striped bass) to 4.2 mm (threadfin shad).  Exact numbers of fish in
each test were unknown but ranged from 10 to over 200.  

The study in Hanson et al. (1977) was performed in an oval test flume
(0.84 m wide and 1.22 m deep) at the Delmarva Power and Light test
laboratories.  While this reference documents several phases of testing
to evaluate entrainment and impingement of eggs and larvae by a profile
wire screen with 1 mm slot width, only a set of tests involving the
introduction of 100 striped bass larvae into the test flume appeared to
measure entrainment through the screen, and therefore, was relevant for
this memorandum.  These tests were performed at an intake velocity
ranging from 0.152 to 0.46 m/s, and duration of 30 minutes.  Larval
lengths ranged from 8 to 17 mm.

Note that only EPRI (2003) controlled for slot velocity in its tests; no
other document reported slot velocity.  Furthermore, EPRI (2008) and
Tomljanovich et al. (1978) reported “percent retention” data, rather
than percent entrained data.  EPRI (2008) also reported the number of
organisms that were released in the test flume.  As noted below, we
assumed that any fish not “retained” was entrained, and therefore,
we calculated the numbers of organisms entrained as the number released
minus the number retained.  Hanson et al. (1977) did not report the
number of organisms released, and therefore, we calculated percent
entrainment as 100% minus percent retention for this study.

When numbers of entrained (or retained) organisms were reported in these
references, they were summed across species and test replicates. 
Likewise, when numbers of organisms released into the test flume were
reported, they were summed in the same manner.  Both sets of sums appear
in Tables 1 through 3.  Percent entrainment was then calculated as the
total number of entrained organisms divided by the total number
released, and the tables include this result.  If a reference only
reported percent entrainment, with no accompanying numbers of organisms
entrained or released, we calculated the average of these percentages
across species and test replicates and presented them in Tables 1
through 3.

Data Acceptance Criteria 

In order to be included in Tables 1 through 3 (and accepted for
statistical analysis), entrainment data needed to achieve the following
criteria:

The data must originate from a study performed at a test laboratory
under controlled conditions (e.g., not in a test barge placed in a
natural water body).

The test screens in this study must be categorized as “fine mesh”
screens, where the mesh size is clearly specified.

From the reported data, it must be possible to determine the percent of
organisms that were entrained through the test screens.  

Data sources reporting only numbers of entrained fish were not accepted
if the total number of fish released was not also reported.

Data were not accepted if numbers of entrained organisms could not be
determined with sufficient confidence from them.  For example, if a data
source reported a number or percentage of fish that avoided either
impingement or entrainment and no other data, then these data were not
accepted for use because it would not be possible to determine only
those fish that were entrained from the reported data.  

While data were reported specifically on entrainment in EPRI (2008) and
Hanson et al. (1977), the studies in EPRI (2003) and Tomljanovich et al.
(1978) measured “retainment” rather than entrainment.  Organisms are
“retained” when they do not pass through the intake screens, meaning
that they are either impinged on the screens, are able to keep swimming
in front of the screens, etc.  Thus, one may be able to assume that all
organisms that are not retained are entrained through the screens.  In
order to accept data for studies that reported only retention data, it
was necessary to make this assumption.  Therefore, when utilizing data
from EPRI (2003) and Tomljanovich et al. (1978) to characterize
entrainment in this memorandum, this assumption should be noted.

Plots of Percent Entrainment Data

Figures 1 through 6 contain plots of the percent entrainment data, with
Figures 1 and 2 presenting data for larvae, Figures 3 and 4 for eggs,
and Figures 5 and 6 for all organisms.  Odd-numbered figures plot the
data versus screen slot width, with slot velocity noted by the plotting
symbol.  Even-numbered figures include data only for EPRI (2003), with
slot velocity specified along the horizontal axis and screen slot width
given by the plotting symbol.  The values plotted in these figures
represent both parallel and perpendicular screen orientations (relative
to flow).

Figure 1.	Percent of Larvae Entrained by Slot Velocity and Screen Size,
with Screen Size on the Horizontal Axis

Figure 2.	Percent of Larvae Entrained by Slot Velocity and Screen Size,
with Slot Velocity on the Horizontal Axis

Figure 3.	Percent of Eggs Entrained by Slot Velocity and Screen Size,
with Screen Size on the Horizontal Axis

 

Figure 4.	Percent of Eggs Entrained by Slot Velocity and Screen Size,
with Slot Velocity on the Horizontal Axis

Figure 5.	Percent of Total Organisms Entrained by Slot Velocity and
Screen Size, with Screen Size on the Horizontal Axis

Figure 6.	Percent of Total Organisms Entrained by Slot Velocity and
Screen Size, with Slot Velocity on the Horizontal Axis

Note that because a large portion of the percent entrainment data
values for eggs are zero (from the EPRI 2001 study), Figures 3 and 4
appear to have fewer plotting points compared to the other plots.  As
noted in Table 2, no surrogate eggs were entrained in the EPRI 2001
study except when the 2.0 mm screens were oriented perpendicular to
flow.  In addition, no entrainment of white sucker eggs occurred in the
EPRI 2002 study (although alewife eggs were entrained at high
percentages).  In both studies, the lack of entrainment was due to the
large egg diameters related to the slot widths and the eggs’
“ability to maintain their shape and not be forced through the
screens” (EPRI, 2003).

These plots suggest that percent entrainment tends to increase with
increased screen slot width for both egg and larval life stages.  In
contrast, increased slot velocity was more highly associated with
increased percent entrainment only for eggs.  (This was observed
primarily due to increased entrainment with alewife eggs at the 0.30 m/s
slot velocity in the EPRI 2002 study.)

Conclusions on Slot Width and Slot Velocity as Noted by the Study
Authors

The following types of conclusions were made by the authors of the above
studies upon review and analysis of the entrainment data.  Only EPRI
(2003) based some of its conclusions on entrainment or retainment data
on the outcome of a statistical analysis (ANOVA).

EPRI (2003)

Entrainment increased with increases in slot size and slot velocity.

Entrainment decreased with increases in channel velocity.

The ability to characterize relationships in entrainment among the
various test factors was hampered by interactions among the factors or
data gaps.

Most larvae and eggs were excluded from passing through 0.5 mm screens,
but not larger mesh screens.

In the 2001 study, under the parallel screen orientation, entrainment
rates increased significantly with slot size (at the 0.05 level) and
were higher at the 0.3 m/s slot velocity.

In the 2002 study, statistical analysis was performed by species.  Slot
velocity effects were significant at the 0.05 level on percent
entrainment for larvae of selected species (e.g., winter flounder,
yellow perch, carp).  While it was not statistically significant, the
analysis on winter flounder data indicated that the impact of slot
velocity was greatest at the smallest screen mesh size.

EPRI (2008)

“The 0.5 mm screens used during survival testing were effective in
retaining the majority of fish released.”

“In general, retention increased as (approach) velocity and duration
of impingement increased.”

Tomljanovich et al. (1978) 

“Velocity and test duration contributed little to percent
retention.”

“Body length alone was not a good indicator of percent retention.”

In addition, EPRI (2003) cites Hanson et al. (1977) in making the
following statement:  “Impingement and entrainment have been
positively correlated with slot velocity and inversely related to
ambient velocity.”  Hanson et al. (1977) noted that over 90% of the
tested larvae were able to swim for 30 minutes without entrainment when
subject to currents of 1.0 to 1.5 fps.

Statistical Analysis of Percent Entrainment Data in Tables 1 through 3

We utilized analysis of variance (ANOVA) methods to investigate the
possible effect of screen size and slot velocity on the percent of
organisms entrained as reported in Tables 1 through 3 above.  We applied
ANOVA separately to each table (i.e., organisms in the egg life stage,
organisms in the larval (non-egg) life stage, and organisms in both life
stages).  We considered several ANOVA model forms that differed
according to whether percent entrainment was transformed, whether slot
velocity was included as a model predictor, and, if slot velocity was
included, whether an interaction term between screen size and slot
velocity should be included.  

Note that for eggs, percent entrainment was reported to be zero for most
test conditions (Table 2), in part due to the limited number of
laboratory studies that considered eggs, and to limited variation in the
different types of eggs that were considered in these studies (EPRI,
2003).  Thus, conclusions from the statistical analyses performed only
on eggs should be made with caution.

βj + (αβ)ij +(ijk					(1)

where Yijk is the “inverse arcsine”-transformed percent entrainment
reported by the kth study under the ith slot width and jth slot velocity
conditions, ( is an overall constant, (i is the amount contributed by
the ith slot width, (j is the amount contributed by the jth slot
velocity, (αβ)ij is the amount contributed by the combination of the
ith slot width and the jth slot velocity, and (ijk is random error left
unexplained by the model.  Because only two values (0.15 m/s and 0.30
m/s) for slot velocity were represented among the available data, we
considered the variable to be categorical like the slot width variable. 


The “inverse arcsine” transformation used in model (1), which EPRI
(2003) also considered in its statistical analysis, is defined as
follows:

,

where Xijk is the percent entrained associated with the kth study that
utilized the ith slot width and the jth slot velocity.  We transformed
the percent entrainment measure because plots of residuals from the
ANOVA model fitting suggested that, for all three subsets of data, the
errors were more likely to be independent and normally distributed with
a constant variance upon taking this transformation.  We also considered
a logarithmic transformation, which appeared to perform similarly to the
inverse arcsine transformation.  

Upon fitting model (1) to data in Tables 1 through 3, we noted the
following:

The interaction effect between screen size and slot velocity was
significant for percent of larvae entrained (p-value = 0.0134) and
percent of eggs entrained (p-value = 0.0198).  Since there is an
interaction effect, the effect of slot velocity varies by screen size.
For example, percent entrainment of larvae appears to increase at the
lower slot velocity at 2.0 mm mesh size.  The least squares estimates of
the mean percent entrainment for each combination of screen size and
slot velocity are shown in Table 4.

Only screen width was significant for percent of all organisms entrained
(p-value = 0.0263).  The smallest screen width (0.5 mm) had the smallest
percent entrainment, while the largest screen width had among the
largest percent entrainment values.  

Table 4.	Least Squares Means and 95% Confidence Limits for Percent
Entrainment Measured in EPRI (2003), by Screen Width and Slot Velocity,
As Estimated from a Two-Way ANOVA Model 

Data Set	Screen Width (mm)	Slot Velocity (m/s)	Least Squares Means	95%
Confidence Limits for Least Squares Means

Larvae Only	0.5	0.15	4.3%	(0.6, 22.5)

	0.5	0.30	10.0%	(0.1, 32.4)

	1.0	0.15	43.1%	(12.9, 76.5)

	1.0	0.30	58.3%	(30.4, 83.6)

	2.0	0.15	80.8%	(54.7, 97.1)

	2.0	0.30	2.7%	(10.4, 37.4)

Eggs Only	0.5	0.15	0.0%	(0.0, 1.7)

	0.5	0.30	0.7%	(0.2, 4.4)

	1.0	0.15	0.0%	(0.0, 5.0)

	1.0	0.30	0.0%	(0.5, 2.5)

	2.0	0.15	0.6%	(0.7, 5.5)

	2.0	0.30	36.2%	(16.6, 58.5)

All Organisms	0.5	0.15	3.0%	(2.0, 22.1)

	0.5	0.30	6.0%	(0.5, 28.6)

	1.0	0.15	34.5%	(5.7, 72.2)

	1.0	0.30	37.9%	(11.6, 68.9)

	2.0	0.15	53.1%	(23.0, 82.0)

	2.0	0.30	22.8%	(0.2, 74.8)



One-way ANOVA.  In order to apply more data from the above four
documents into the ANOVA, we also fit a second model that included only
screen mesh width as a predictor variable.  Thus, this model could
assess only the effect of screen mesh width.  The form of this model is:

  Yjk = ( + (i + (jk 					(2)

where the notation is the same as for model (1).  We again investigated
the residuals for this model and concluded that the assumptions of the
ANOVA method were satisfied.  Based on the results of the three runs of
the model, we noted the following:

Screen width had a significant effect on average percent of larvae
entrained (p-value = 0.002) and the average percent of total organisms
entrained (p-value = 0.0109).  This appears to be due to low percent
entrainment outcomes associated with the 0.5 mm mesh size.  Greater
percentages of entrainment were associated with larger mesh sizes.

Screen width alone did not have a significant effect on the average
percent of eggs entrained.

The least squares estimates of the mean effect of screen width are
reported in Table 5.

Table 5.	Least Squares Means and 95% Confidence Limits for Percent
Entrainment Measured in Controlled Laboratory Studies, by Screen Width,
As Estimated from a One-Way ANOVA model with Screen Width as a Factor  

Data Set	Screen Width (mm)	Least Squares Means	95% Confidence Limits for
Least Squares Means

Larvae Only	0.5	5.2%	(0.2, 16.3)

	1.0	41.7%	(21.4, 63.6)

	1.3	35.7%	(0.3, 88.6)

	1.8	20.2%	(1.4, 75.4)

	2.0	59.2%	(30.6, 84.8)

	2.5	32.6%	(0.0, 86.4)

Eggs Only	0.5	0.2%	(1.5, 4.0)

	1.0	0.0%	(0.0, 5.2)

	2.0	6.9%	(0.1, 22.6)

All Organisms	0.5	3.8%	(0.4, 10.4)

	1.0	30.7%	(17.5, 46.1)

	1.3	35.7%	(4.9, 75.8)

	1.8	20.2%	(0.2, 59.6)

	2.0	45.1%	(25.4, 65.6)

	2.5	32.6%	(3.6, 73.0)



The outcome of the above statistical analyses were compared to that
observed with percent reduction in entrainment from field studies, as
reported in Battelle’s technical memorandum dated September 29, 2008. 
The above analysis on laboratory data noted significant slot velocity
effects (at the 0.05 level) separately for eggs and larvae (but not for
all organisms) which varied according to the slot width size.  However,
as noted in the 9/29 memo, no significant slot velocity effects were
observed in percent reduction in entrainment in field studies for any
life stage, and the interaction of slot velocity and slot width was not
significant.  

In both lab and field studies, increased entrainment was observed to
occur with increased mesh sizes.  However, the outcome of the
statistical analyses (involving the one-way ANOVA) differed between the
two types of studies.  While screen mesh size was observed to have a
significant effect on percent reduction in entrainment within the field
studies for eggs only (and not for larvae or total organisms), mesh size
was a significant factor for percent entrainment in the lab studies for
larvae and total organisms (but not for eggs).  For eggs, greater
entrainment levels were observed in the field studies compared to the
lab studies documented in EPRI (2003).  Again, caution should be made
when comparing data outcomes between the field and lab studies due to
differences in how the performance measure was defined and in the ANOVA
models applied to the performance measure data.

References

EPRI (2008).  Laboratory Evaluation of Fine-Mesh Traveling Water Screens
for Protecting Early Life Stages of Fish at Cooling Water Intakes.  Palo
Alto, CA:  Electric Power Research Institute, Inc.  Technical Report No.
1014021.  March 2008.

EPRI (2003).  Laboratory Evaluation of Wedgewire Screens for Protecting
Early Life Stages of Fish at Cooling Water Intakes.  Palo Alto, CA: 
Electric Power Research Institute, Inc.  Technical Report No. 1005339. 
May 2003.

Hanson, BN, Bason, WH, Beitz, BE, and Charles, KE. (1977).  A practical
intake screen which substantially reduces the entrainment and
impingement of early life stages of fish.  In:  Fourth National Workshop
on Entrainment and Impingement.  LD Jenson, ed.  Melville, NY:  EA
Communications (a division of Ecological Analysts, Inc.)

Tomljanovich, DA, Heuer, JH, and Voighlander, CW.  (1978). 
Investigations on the protection of fish larvae at water intakes using
fine-mesh screening.  In:  Larval Exclusion Systems for Power Plant
Cooling Water Intakes:  Proceedings of the Workshop Held at Shelter
Island Inn, San Diego, CA, February 7-8, 1978.  RK Sharma and JB Palmer,
eds.  Argonne, IL:  Argonne National Laboratory.

October 17, 2008

316b:  Laboratory Test Data Related to Entrainment (EP-C-05-030; WA
3-13)

 h

 h

h

 h

h8

h8

h8

h8

h

h8

h8

h8

h8

h8

@

@

@

@

g

h

¤

¥

	

-

,

/

f

h

s

£

¤

¥

¦

h8

  h

h8

 h

h

 h

 h

 h

 h

 h

¦

®

±

ã

?

@

w

À

Á

Å

Ü

æ

ï

h

h

h!

h!

h!

hë

hë

 hë

 hë

hë

hë

hë

hë

hë

 hë

hë

  hë

hë

hë

hë

h8

@

@

  PAGE  \* MERGEFORMAT  15 

October 17, 2008

316b:  Laboratory Test Data Related to Entrainment (EP-C-05-030; WA
3-13)

Page   PAGE  \* MERGEFORMAT  2 

