MEMORANDUM

Tetra Tech, Inc.

400 Red Brook Blvd., Suite 200

Owings Mills, MD 21117

phone	410-356-8993

fax	410-356-9005

DATE: 		March 20, 2008

TO:			Paul Shriner, EPA

	

FROM:		Henry Latimer, Blaine Snyder, and Chris Millard

SUBJECT:		Taxonomically-based fish species grouping scheme

As part of a larger effort to interpret potential impingement mortality
and entrainment (IM&E) effects, it is desirable to organize fish species
into groups with similar IM&E vulnerabilities.  Ideally, these groups
would comprise fish species with similar characteristics (e.g., swim
speed, schooling behavior, reproductive and life history
characteristics) as well as habitat preferences (e.g., benthic vs.
pelagic, river vs. lake, freshwater vs. estuarine).  However, given the
number of fish species potentially exposed to IM&E and the effort
required to summarize relevant information to be used in grouping these
species, a first effort based simply on taxonomy has been undertaken. 
The goals of this effort are to 1) help define the scope of the species
grouping effort and 2) better understand the types of information
required to successfully assemble species into meaningful groups. 

There are currently 3,694 fish species in North America (Nelson et al.
2004).  Functionally, many of these species are not likely to be exposed
to cooling water intake structures (CWIS) in the U.S., either because
they occur only in Mexico or Canada or are pelagic marine species that
are unlikely to occur in near-shore waters.  The list of North American
fish species was screened to eliminate those that do not occur in the
U.S. and those that are unlikely to occur in near-shore waters. 
Following this screening, 1,279 species remain that could reasonably be
expected to occur in the U.S. in the vicinity of CWISs and be exposed to
IM&E related to those CWISs.  This list includes all freshwater and most
(if not all) estuarine fish.  Future, comprehensive species grouping
efforts will require that marine species be reviewed in greater depth to
determine which species should be considered to be reasonably vulnerable
to IM&E.

To minimize the number of fish species groups, the taxonomic level
selected for use as a basis for groupings must be fairly general, but
specific enough to capture meaningful morphological and habitat-specific
differences among the various species.  For this exercise, we selected
Family as the appropriate level for use as the basis in establishing
groups.  Dividing the groups based on Class was determined to result in
species groupings that lumped clearly dissimilar species together. 
Groups based on Genera would be preferable in terms of lumping more
similar species together, however such an approach yields so many groups
(278) as to limit the utility of a grouping approach.

The family-level evaluation yields 72 different groups (see attached
spreadsheet).  It is likely that in refining this approach, similarities
among some Families will be identified that will allow for several of
these distinct groups to be combined.  However, it is also known that
variations in morphology and preferred habitat within some Families will
require that those groupings be further sub-divided. 

Based on a brief analysis, it is possible that some species from
differing families could be combined.  For example, some members of the
sculpin Family (Cottidae) would be able to be combined with some members
of the Perch Family (e.g., darters), in that some species within each
group exhibit similar, benthic habitat preferences.  However, further
investigation into habitat preferences, geographic ranges, and swim
speed analyses is required before such groupings can be made.  It is
also likely that some, more diverse Families would require further
division to more accurately represent the IM&E vulnerabilities of these
groups.  For instance, the minnow family (Cyprinidae) which contains
carp, minnows, and pike minnows would likely have to be divided into at
least three groups that are reflective of the diverse body sizes,
habitats, and swim speeds of fish within this family.  

Thus, while a taxonomically-based species grouping effort is a
reasonable initial step in grouping fish species in terms of IM&E
vulnerability, there are clear shortcomings of this approach.  To
assemble larger and more inclusive species groups, the scientific
literature must be evaluated to gather data on species-specific habitat
preferences, geographic range, and swim speeds.  Such information is
required to determine whether the groupings based on potential IM&E
vulnerability are similar to existing taxonomic groupings and, if not,
how such groupings could be refined.  Further, information should be
assembled for more than one life-stage to ensure that the IM&E
vulnerability of potentially more sensitive (but less well-studied) life
stages (e.g., egg) are adequately addressed.  It is likely that the
scientific literature does not contain complete, relevant information
for the majority (possibly the vast majority) of species.  It is
expected that this more in depth approach would be able to propose IM&E
vulnerability values for a subset of all fish species.  These values
would then be extrapolated and applied to all taxonomically similar
species within each group.  The scientific validity of the final
proposed grouping schemes would be directly dependent upon the amount of
available species-specific data available in the literature.  

An alternative approach would be to refine the species list (currently
1,279 for freshwater and estuarine in the U.S. and 3,694 if all North
American fish are included) based upon those fish that can reasonably be
expected to be subject to IM&E.  A judgment as to which species can
reasonably be expected to be subject to IM&E could be made based upon an
analysis of those fish species actually impinged and entrained through
CWISs at representative facilities (e.g., encompassing intake types,
sizes, and locations) throughout the U.S.  Such data from facilities
located in all geographic (e.g., Northeast, Southwest, Midwest) areas
and water body types (e.g., river, lake, Great Lakes, estuary, marine)
have already been collected and are available in state and/or EPA
records.  Evaluation of assembled impingement and entrainment (I&E) data
would be expected to yield a much more limited list of species which are
demonstrably subject to IM&E in the U.S.  

A list of species that are actually subject to IM&E would be useful for
a variety of reasons.  This list would be expected to represent a small
percentage of all species in the U.S., thus enabling species grouping
efforts to be streamlined, more relevant, and likely more scientifically
defensible.  Such a list would also reveal real-world data that is
directly relevant to species specific IM&E vulnerability which may be
more useful that literature-derived data that may have questionable
application in determining IM&E vulnerability.  The resulting list would
also allow alternative approaches to be explored to evaluate and
regulate IM&E.  Further, this list would allow EPA to make more informed
decisions as to how IM&E regulations should be applied to various water
body types and how the various water body types could reasonably be
combined (e.g., address whether separation of lakes and rivers is
appropriate or necessary).

We are optimistic that some type of species grouping scheme is possible.
 However, we believe that inclusion of actual I&E data is a necessary
step to screen the list of all possible fish species to assemble a list
of fish species that can reasonably be expected to be subject to IM&E. 
Using this revised list, a more appropriate and defensible IM&E
vulnerability assessment could be made of each species, and reasonable
species groupings could then be assembled with a fairly high degree of
confidence.

Literature Cited

Nelson, J.S., E.J. Crossman, H. Espinosa-Perez, L.T. Findley, C.R.
Gilbert, R.N. Lea, and J.D. Williams.  2004.  Common and scientific
names of fishes from the Unites States, Canada, and Mexico.  American
Fisheries Society, Special Publication 29, Bethesda, Maryland.

