Endangered and Threatened Species; Determination of Endangered Status for the Gulf of Maine Distinct Population Segment of Atlantic Salmon, 29344-29387 [E9-14269]

Agencies

• DEPARTMENT OF COMMERCE
• National Oceanic and Atmospheric Administration
• Fish and Wildlife Service
• Department of Interior

[Federal Register Volume 74, Number 117 (Friday, June 19, 2009)]
[Rules and Regulations]
[Pages 29344-29387]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: E9-14269]



[[Page 29343]]

-----------------------------------------------------------------------

Part III





Department of the Interior





-----------------------------------------------------------------------



Fish and Wildlife Service



-----------------------------------------------------------------------





Department of Commerce





-----------------------------------------------------------------------



National Oceanic and Atmospheric Administration



-----------------------------------------------------------------------



50 CFR Parts 17 and 224



Endangered and Threatened Species; Determination of Endangered Status 
for the Gulf of Maine Distinct Population Segment of Atlantic Salmon; 
Final Rule

Federal Register / Vol. 74, No. 117 / Friday, June 19, 2009 / Rules 
and Regulations

[[Page 29344]]


-----------------------------------------------------------------------

DEPARTMENT OF INTERIOR

Fish and Wildlife Service

50 CFR Part 17

DEPARTMENT OF COMMERCE

National Oceanic and Atmospheric Administration

50 CFR Part 224

[Docket No. 0808191116-9709-02]
RIN 0648-XJ93


Endangered and Threatened Species; Determination of Endangered 
Status for the Gulf of Maine Distinct Population Segment of Atlantic 
Salmon

AGENCY: National Marine Fisheries Service (NMFS), National Oceanic and 
Atmospheric Administration (NOAA), Commerce; United States Fish and 
Wildlife Service (USFWS), Interior.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: We (NMFS and USFWS, collectively referred to as the Services) 
have determined that naturally spawned and conservation hatchery 
populations of anadromous Atlantic salmon (Salmo salar) whose 
freshwater range occurs in the watersheds from the Androscoggin River 
northward along the Maine coast to the Dennys River, including those 
that were already listed in November 2000, constitute a distinct 
population segment (DPS) and hence a ``species'' for listing. We have 
determined that the Gulf of Maine (GOM) DPS warrants listing as 
endangered under the Endangered Species Act (ESA). Critical habitat for 
the GOM DPS will be designated in a subsequent Federal Register notice.

DATES: This rule is effective July 20, 2009.

ADDRESSES: Comments and materials received, as well as supporting 
scientific information used in the preparation of this rule, will be 
available for public inspection, by appointment, during normal business 
hours at: National Marine Fisheries Service, Northeast Regional Office, 
55 Great Republic Drive, Gloucester MA 01930. An electronic copy of 
this final rule is available at: https://www.nero.noaa.gov/prot_res/altsalmon/. Public comments received can be viewed at https://www.regulations.gov.

FOR FURTHER INFORMATION CONTACT: Rory Saunders, NMFS, at (207) 866-
4049; Jessica Pruden, NMFS, at (978) 282-8482; Marta Nammack, NMFS, at 
(301) 713-1401; Lori Nordstrom, USFWS, at (207) 827-5938 ext. 13. 
Persons who use a Telecommunications device for the deaf (TDD) may call 
the Federal Information Relay Service (FIRS) at 1-800-877-8339, 24 
hours a day, 7 days a week.

SUPPLEMENTARY INFORMATION:

Background

    We issued a final rule listing the GOM DPS of Atlantic salmon as 
endangered on November 17, 2000 (65 FR 69469). The GOM DPS was defined 
as all naturally reproducing wild populations and those river-specific 
hatchery populations of Atlantic salmon having historical, river-
specific characteristics found north of and including tributaries of 
the lower Kennebec River to, but not including, the mouth of the St. 
Croix River at the U.S.-Canada border. In the final rule listing the 
GOM DPS, we did not include fish that inhabit the mainstem and 
tributaries of the Penobscot River above the site of the former Bangor 
Dam, the upper Kennebec River, or the Androscoggin River within the GOM 
DPS (65 FR 69469; November 17, 2000).
    In late 2003, we assembled the 2005 Biological Review Team (BRT) 
composed of biologists from the Maine Atlantic Salmon Commission (now 
the Maine Department of Marine Resources Bureau of Sea-run Fisheries 
and Habitat (MDMR)), the Penobscot Indian Nation, and both Services. 
The 2005 BRT was charged with reviewing and evaluating all relevant 
scientific information relating to the current DPS delineation 
(including a detailed genetic characterization of the Penobscot 
population and data relevant to the appropriateness of including the 
upper Kennebec and Androscoggin rivers as part of the DPS), determining 
the conservation status of the populations not included in GOM DPS 
listed in 2000, and assessing their relationship to the GOM DPS as it 
was listed in 2000. The findings of the 2005 BRT, which are detailed in 
the 2006 Status Review for Anadromous Atlantic Salmon in the United 
States (Fay et al., 2006), addressed: the DPS delineation, including 
whether populations that were not included in the 2000 listing should 
be included in the GOM DPS; the extinction risks to the species; and 
the threats to the species. The 2006 Status Review (Fay et al., 2006) 
underwent peer review by experts in the fields of Atlantic salmon 
biology and genetics to ensure that it was based on the best available 
science. Each peer reviewer independently affirmed the major 
conclusions presented in Fay et al. (2006).

Policies for Delineating Species Under the ESA

    Section 3 of the ESA defines ``species'' as including ``any 
subspecies of fish or wildlife or plants, and any distinct population 
segment of any species of vertebrate fish or wildlife which interbreeds 
when mature.'' The term ``distinct population segment'' is not 
recognized in the scientific literature. Therefore, the Services 
adopted a joint policy for recognizing DPSs under the ESA (DPS Policy; 
61 FR 4722) on February 7, 1996. The DPS policy requires the 
consideration of two elements when evaluating whether a vertebrate 
population segment may be considered a DPS under the ESA: (1) The 
discreteness of the population segment in relation to the remainder of 
the species or subspecies to which it belongs; and (2) the significance 
of the population segment to the species or subspecies to which it 
belongs.
    A population segment of a vertebrate species may be considered 
discrete if it satisfies either one of the following conditions: (1) It 
is markedly separated from other populations of the same taxon (an 
organism or group of organisms) as a consequence of physical, 
physiological, ecological, or behavioral factors. Quantitative measures 
of genetic or morphological discontinuity may provide evidence of this 
separation; or (2) it is delimited by international governmental 
boundaries within which differences in control of exploitation, 
management of habitat, conservation status, or regulatory mechanisms 
exist that are significant in light of section 4(a)(1)(D) of the ESA 
(i.e., inadequate regulatory mechanisms).
    If a population segment is found to be discrete under one or more 
of the above conditions, its biological and ecological significance to 
the taxon to which it belongs is evaluated. This consideration may 
include, but is not limited to: (1) Persistence of the discrete 
population segment in an ecological setting unusual or unique for the 
taxon; (2) evidence that the loss of the discrete population segment 
would result in a significant gap in the range of a taxon; (3) evidence 
that the discrete population segment represents the only surviving 
natural occurrence of a taxon that may be more abundant elsewhere as an 
introduced population outside its historic range; and (4) evidence that 
the discrete population segment differs markedly from other populations 
of the species in its genetic characteristics.

[[Page 29345]]

Listing Determinations Under the ESA

    The ESA defines an endangered species as one that is in danger of 
extinction throughout all or a significant portion of its range, and a 
threatened species as one that is likely to become endangered in the 
foreseeable future throughout all or a significant portion of its range 
(sections 3(6) and 3(20), respectively). The statute requires us to 
determine whether any species is endangered or threatened because of 
any of the following five factors: (1) The present or threatened 
destruction, modification, or curtailment of its habitat or range; (2) 
overutilization for commercial, recreational, scientific, or 
educational purposes; (3) disease or predation; (4) the inadequacy of 
existing regulatory mechanisms; or (5) other natural or manmade factors 
affecting its continued existence (section 4(a)(1)(A-E)). We are to 
make this determination based solely on the best available scientific 
and commercial data available after conducting a review of the status 
of the species and taking into account any efforts being made by states 
or foreign governments to protect the species.

Atlantic Salmon Life History

    Anadromous Atlantic salmon are a wide ranging species with a 
complex life history. The historic range of Atlantic salmon occurred on 
both sides of the North Atlantic: from Connecticut to Ungava Bay in the 
western Atlantic and from Portugal to Russia's White Sea in the Eastern 
Atlantic, including the Baltic Sea.
    For Atlantic salmon in the United States, juveniles typically spend 
2 years rearing in freshwater. Freshwater ecosystems provide spawning 
habitat and thermal refuge for adult Atlantic salmon; overwintering and 
rearing areas for eggs, fry, and parr; and migration corridors for 
smolts and adults (Bardonnet and Bagliniere, 2000). Adult Atlantic 
salmon typically spawn in early November. During spawning, the female 
uses its tail to scour or dig a series of nests in the gravel where the 
eggs are deposited; this series of nests is called a redd. The eggs 
remain in the redd until they hatch in late March or April. At this 
stage, they are referred to as alevin or sac fry. Alevins remain in the 
redd for about 6 more weeks and are nourished by their yolk sac until 
they emerge from the gravel in mid-May. At this time, they begin active 
feeding and are termed fry. Within days, the fry enter the parr stage, 
indicated by vertical bars (parr marks) on their sides that act as 
camouflage. Atlantic salmon parr are territorial; thus, most juvenile 
mortality is thought to be density dependent and mediated by habitat 
limitation (Gee et al., 1978; Legault, 2005). In particular, suitable 
overwintering habitat may limit the abundance of large parr prior to 
smoltification (Cunjak et al., 1998). Smoltification (the physiological 
and behavioral changes required for the transition to salt water) 
usually occurs at age 2 for most Atlantic salmon in Maine. The smolt 
emigration period is rather short and lasts only 2 to 3 weeks for each 
individual. During this brief emigration window, smolts must contend 
with rapidly changing osmoregulatory requirements (McCormick et al., 
1998) and predator assemblages (Mather, 1998). The freshwater stages in 
the life cycle of the Atlantic salmon have been well studied; however, 
much less information is available on Atlantic salmon at sea (Klemetsen 
et al., 2003).
    Gulf of Maine Atlantic salmon migrate vast distances in the open 
ocean to reach feeding areas in the Davis Strait between Labrador and 
Greenland, a distance over 4,000 km from their natal rivers (Danie et 
al., 1984; Meister, 1984). During their time at sea, Atlantic salmon 
undergo a period of rapid growth until they reach maturity and return 
to their natal river. Most Atlantic salmon (about 90 percent) from the 
Gulf of Maine return after spending 2 winters at sea; usually less than 
ten percent return after spending 1 winter at sea; roughly one percent 
of returning salmon are either repeat spawners or have spent 3 winters 
at sea (3 sea winter, or 3SW salmon) (Baum, 1997).
    In addition to anadromous Atlantic salmon, landlocked Atlantic 
salmon have been introduced to many lakes and rivers in Maine, though 
they are only native to four watersheds in the State: The Union, 
including Green Lake in Hancock County; the St. Croix, including West 
Grand Lake in Washington County; the Presumpscot, including Sebago Lake 
in Cumberland County; and the Penobscot, including Sebec Lake in 
Piscataquis County (Warner and Havey, 1985). There are certain lakes 
and rivers in Maine where landlocked salmon and anadromous salmon co-
exist. Recent genetic surveys have confirmed that little genetic 
exchange occurs between these two life history types (Spidle et al., 
2003; NMFS unpublished data).

Delineation of the Gulf of Maine Distinct Population Segment

    Fay et al. (2006) concluded that the DPS delineation that resulted 
in the 2000 listing designation (65 FR 69469; November 17, 2000) was 
largely appropriate, except in the case of large rivers that were 
excluded in the previous listing determination (Section 6.2.4 of Fay et 
al., 2006). As described below in the analyses of discreteness and 
significance of the population segment, Fay et al. (2006) concluded 
that the salmon currently inhabiting the larger rivers (Androscoggin, 
Kennebec, and Penobscot) are genetically similar to the rivers included 
in the GOM DPS as listed in 2000 (Spidle et al., 2003), have similar 
life history characteristics, and occur in the same zoogeographic 
region (section 6.3 of Fay et al., 2006). Further, the salmon 
populations inhabiting the large and small rivers from the Androscoggin 
River northward to the Dennys River differ genetically and in important 
life history characteristics from Atlantic salmon in adjacent portions 
of Canada (Spidle et al., 2003; Fay et al., 2006). Thus, Fay et al. 
(2006) (section 6.3.1.4 and 6.3.2.4) concluded that this group of 
populations (population segment) met both the discreteness and 
significance criteria of the DPS Policy and, therefore should be 
considered a DPS. Fay et al. (2006) recommended that the new GOM DPS 
include all anadromous Atlantic salmon whose freshwater range occurs in 
the watersheds from the Androscoggin River northward along the Maine 
coast to the Dennys River, including all associated conservation 
hatchery populations used to supplement these natural populations; 
currently, such conservation hatchery populations are maintained at 
Green Lake National Fish Hatchery (GLNFH) and Craig Brook National Fish 
Hatchery (CBNFH).

Delineating Geographic Boundaries

    Determining the precise boundary of the GOM DPS is difficult. In 
the case of the GOM DPS, we use a wide array of independent sources of 
information to make this determination. These sources of information 
include recent genetic analyses, life history, and zoogeography, among 
others. Recent genetic analyses, in particular, have clarified these 
distinctions, and we rely on them heavily in the following analysis. 
When using genetic data to make these delineations, it is important to 
note that extant populations must exist in order to make meaningful 
comparisons. In the case of determining the northern boundary of the 
GOM DPS, extant populations were used in genetic analyses and thus 
inform the determination. However, in the case of the determination of 
the southern boundary of the GOM DPS, many populations south of the 
Androscoggin are extirpated, and thus there are no genetic data 
available to make these

[[Page 29346]]

comparisons. For this reason we rely on additional information to 
delineate the southern boundary of the GOM DPS below.
    We relied on genetic data to inform our determination on the 
northern terminus of the GOM DPS. At a broad scale, it is clear that 
there are substantial differences in genetic structure between U.S. and 
Canadian populations of Atlantic salmon (Spidle et al., 2003). However, 
there are no genetic data on the wild salmon that once occurred in the 
St. Croix watershed along the U.S.-Canada border. As listed in 2000, 
the northern terminus of the GOM DPS was the U.S.-Canada border at the 
St. Croix River, but as described on page 54 of Fay et al. (2006), the 
best available science suggests that the St. Croix groups with other 
Canadian rivers. Genetic analyses found that salmon in the Dennys River 
are more similar to populations in the United States than to Canadian 
salmon populations that are geographically proximate to the Dennys 
(Spidle et al., 2003). Therefore, we find that the northern terminus of 
the GOM DPS is the Dennys River watershed, rather than the St. Croix.
    We determined the southern terminus of the GOM DPS to be the 
Androscoggin River based on zoogeography rather than genetics because 
there are extremely few Atlantic salmon in the rivers on which to base 
genetic analyses as one moves southward. Due to the combination of low 
numbers of Atlantic salmon in some rivers (e.g., Androscoggin) and the 
complete extirpation of the native stock in other rivers to the south 
(e.g., Merrimack), complete genetic data are not and may never be 
available for the Services to be able to genetically characterize these 
populations. In the absence of clear genetic data, we used ecological 
factors to define the southern boundary of the GOM DPS. The 
Androscoggin River lies within the Penobscot-Kennebec-Androscoggin 
Ecological Drainage Unit (EDU) (Olivero, 2003) and the Laurentian Mixed 
Forest Province (Bailey, 1995), which separates it from more southern 
rivers that were historically occupied by Atlantic salmon. EDUs are 
aggregations of watersheds with similar zoogeographic history, 
physiographic conditions, climatic characteristics, and basic geography 
(Olivero, 2003). The substantial changes in physiographic conditions 
south of the Androscoggin drainage are reflected in the southern 
terminus of both the Laurentian Mixed Forest Province and the 
Penobscot--Kennebec--Androscoggin EDU occurring in that area. Basin 
geography, climate, groundwater temperatures, hydrography, and 
zoogeographic differences between the Penobscot--Kennebec--Androscoggin 
EDU and the EDUs to the south (e.g., Saco-Merrimack-Charles, Lower 
Connecticut, Middle Connecticut, and Upper Connecticut) likely had a 
strong effect upon Atlantic salmon ecology and production. These 
differences would influence the structure and function of aquatic 
ecosystems (Vannote et al., 1980; Cushing et al., 1983; Minshall et 
al., 1983; Cummins et al., 1984; Minshall et al., 1985; Waters, 1995) 
and create a different environment for the development of local 
adaptations than rivers, such as the Saco and Merrimack, to the south.
    In the proposed rule, we proposed to include the entire 
Androscoggin, Kennebec, and Penobscot Watersheds within the GOM DPS 
boundary. Some comments from the public appropriately highlighted 
several impassable falls that limited the upstream extent to which 
anadromous salmon inhabited the rivers of Maine. NMFS also evaluated 
historical occupancy at the watershed scale for the process of 
proposing critical habitat for the GOM DPS. There is also considerable 
information provided in the 2006 Status Review pertaining to impassable 
falls as well. We are, therefore, using these information sources (and 
others cited therein) to delimit the upstream extent of anadromy for 
GOM salmon in this final rule.
    We have identified seven impassable falls that substantially 
limited the upstream extent of the freshwater range of GOM salmon. 
These include Rumford Falls in the town of Rumford on the Androscoggin 
River, Snow Falls in the town of West Paris on the Little Androscoggin 
River, Grand Falls in Township 3 Range 4 BKP WKR, on the Dead River in 
the Kennebec Basin; the un-named falls (impounded by Indian Pond Dam) 
immediately above the Kennebec River Gorge in the town of Indian Stream 
Township on the Kennebec River; Big Niagara Falls on Nesowadnehunk 
Stream in Township 3 Range 10 WELS in the Penobscot Basin; Grand Pitch 
Falls on Webster Brook in Trout Brook Township in the Penobscot Basin; 
and Grand Falls on the Passadumkeag River in Grand Falls Township in 
the Penobscot Basin (Table 1).

         Table 1--Impassable Falls That Limit the Upstream Extent of the Freshwater Range of GOM Salmon
----------------------------------------------------------------------------------------------------------------
            Name of falls                       Town                    River                    Basin
----------------------------------------------------------------------------------------------------------------
Rumford Falls.......................  Rumford................  Androscoggin River....  Androscoggin.
Snow Falls..........................  West Paris.............  Little Androscoggin     Androscoggin.
                                                                River.
Grand Falls.........................  Township 3 Range 4 BKP   Dead River............  Kennebec.
                                       WKR.
Un-named............................  Indian Stream Township.  Kennebec River........  Kennebec.
Big Niagara Falls...................  Township 3 Range 10      Nesowadnehunk Stream..  Penobscot.
                                       WELS.
Grand Pitch.........................  Trout Brook Township...  Webster Brook.........  Penobscot.
Grand Falls.........................  Grand Falls Township...  Passadumkeag River....  Penobscot.
----------------------------------------------------------------------------------------------------------------

    As a result, we have modified the geographic boundaries of the 
freshwater range of GOM salmon in the Androscoggin, Kennebec, and 
Penobscot Basins in the following ways: all freshwater bodies in the 
Androscoggin Basin are included up to Rumford Falls on the Androscoggin 
River and up to Snow Falls on the Little Androscoggin River; all 
freshwater bodies in the Kennebec Basin are included up to Grand Falls 
on the Dead River and the unnamed falls (currently impounded by Indian 
Pond Dam) immediately above the Kennebec River Gorge; and all 
freshwater bodies in the Penobscot Basin are included up to Big Niagara 
Falls on Nesowadnehunk Stream, Grand Pitch on Webster Brook, and Grand 
Falls on the Passadumkeag River.
    We recognize that many other potentially impassable waterfalls 
exist throughout the range of GOM salmon. While other impassable falls 
may exist throughout the range, we did not exclude any other areas 
(other than the areas above the seven falls mentioned above) for the 
following reasons: (1) Their occurrence is typically in headwater areas 
that preclude access from relatively small portions of a given 
watershed; (2) identifying every impassable falls is impractical given

[[Page 29347]]

current information; and (3) no other impassable falls were brought to 
our attention during the public comment period.
    In addition, we recognize that within every watershed, there is an 
upstream extent of anadromy. However, it is impossible to define that 
specific point in every watershed. The upstream extent of anadromy is 
ultimately limited by the incremental narrowing of a given river or 
stream. While a stream may be too small for an adult salmon to swim up 
any further, juveniles may ascend further than that point in search of 
suitable rearing habitat. In fact, upstream movement of even fry can be 
quite substantial. As such, we include all the freshwater bodies as 
part of the freshwater range of GOM salmon unless above one of the 
impassable falls mentioned in the text above.

Discreteness and Significance of the GOM DPS

    With respect to the ``discreteness'' of this population segment, 
section 6.3.1 of Fay et al. (2006) considered ecological, behavioral, 
and genetic factors under the first discreteness criterion of the DPS 
Policy to examine the degree to which it is separate from other 
Atlantic salmon populations. Gulf of Maine salmon are behaviorally and 
physiologically discrete from other members of the taxon because they 
return to their natal GOM rivers to spawn (a process called homing), 
which leads to the separation in stocks that has been observed between 
the Gulf of Maine and other segments of the taxon. River-specific 
adaptation is an important mechanism that allows anadromous salmon to 
occupy diverse environments throughout their range. River-specific 
adaptation is facilitated by homing and is characteristic of all other 
anadromous salmonids (Klemetsen et al., 2003; Utter et al., 2004). Baum 
and Spencer (1990) found that roughly 98 percent of all tagged salmon 
returned to their natal rivers to spawn. As described below, these 
strong homing tendencies have led to the formation and maintenance of 
river-specific adaptations for GOM salmon as well.
    Ecologically, GOM salmon are discrete from other members of the 
taxon. The core of the riverine habitat of this population segment lies 
within the Penobscot-Kennebec-Androscoggin EDU (Olivero, 2003) and the 
Laurentian Mixed Forest Province (Bailey, 1995). These environmental 
conditions have shaped life history characteristics of GOM salmon. In 
particular, GOM salmon life history strategies are dominated by age 2 
smolts and 2SW adults, whereas populations to the north of this 
population segment are generally dominated by age 3 or older smolts and 
1SW adults (called grilse). Smolt age reflects growth rate (Klemetsen 
et al., 2003), with faster growing parr emigrating as smolts earlier 
than slower growing ones (Metcalfe et al., 1990). Smolt age is largely 
influenced by temperature (Symons, 1979; Forseth et al., 2001) and can 
therefore be used to compare and contrast growing conditions across 
rivers (Metcalfe and Thorpe, 1990). For GOM populations, smolt ages are 
quite similar across rivers with naturally-reared (result of either 
wild spawning or fry stocking) returning adults predominantly 
emigrating at river age 2 (88 to 100 percent) with the remainder 
emigrating at river age 3 (Fay et al., 2006). Smolt ages from 
naturally-reared returning adults in rivers south of the Penobscot-
Kennebec-Androscoggin EDU are also dominated by river age 2 smolts with 
some emigrating at river age 3, but a substantial proportion of river 
age 1 smolts are also present (See Table 6.3.1.1 in Fay et al., 2006).
    The strongest evidence that GOM salmon are discrete from other 
members of the taxon is genetic. Fay et al. (2006) described genetic 
structure of this population segment and other stocks in detail in 
section 6.3.1.3. In summary, three primary genetic groups of North 
American populations (Spidle et al., 2003; Spidle et al., 2004; 
Verspoor et al., 2005) are evident. These include the anadromous GOM 
populations (including salmon in the Kennebec and Penobscot Rivers) 
(Spidle et al., 2003), non-anadromous Maine populations (Spidle et al., 
2003), and Canadian populations (Verspoor et al., 2005). Because of 
these behavioral, physiological, ecological and genetic factors, we 
conclude that the GOM anadromous population is discrete from other 
Atlantic salmon populations under the provisions of the DPS Policy.
    With respect to the ``significance'' of this population segment, 
Fay et al. (2006) found that there are three attributes which are 
described as evidence for ``significance'' in the DPS policy that are 
applicable to the GOM DPS (section 6.3.2 of Fay et al., 2006). Fay et 
al. (2006) (section 6.3.2.1) concluded that this population segment has 
persisted in an ecological setting unusual or unique to the taxon for 
several reasons. First, GOM salmon live in and migrate through a unique 
marine environment. The marine migration corridor for GOM salmon begins 
in the GOM that is known for unique circulation patterns, thermal 
regimes, and predator assemblages (Townsend et al., 2006). Gulf of 
Maine salmon undertake extremely long marine migrations to feeding 
grounds off the West Coast of Greenland because the riverine habitat 
they occupy is at the southern extreme of the current North American 
range. While such vast marine migrations are more common for stocks on 
the northeast side of the Atlantic, the combination of the long 
migration distances and the unique setting of the GOM, described above, 
make the oceanic life history of the GOM DPS quite different from those 
of other stocks (ICES, 2008). In addition, the core of the riverine 
habitat of this population segment lies within the Penobscot-Kennebec-
Androscoggin EDU (Olivero, 2003) and the Laurentian Mixed Forest 
Province (Bailey, 1995). The importance of this setting is evidenced by 
the tremendous production potential of its juvenile nursery habitat 
that allows production of proportionately younger smolts than Canadian 
rivers to the north (Myers, 1986; Baum, 1997; Hutchings and Jones, 
1998). Thus, the combination of the unique rearing conditions in the 
freshwater portion of its range combined with the unique marine 
migration corridor led Fay et al. (2006) to conclude that this 
population segment has persisted in an ecological setting unusual or 
unique to the taxon.
    Fay et al. (2006) also concluded that the loss of this population 
segment would result in a significant gap or constriction in the range 
of the taxon (Section 6.3.2.2 of Fay et al., 2006). The extirpation of 
this population segment would represent a significant range reduction 
for the entire taxon Salmo salar because this population segment 
represents the southernmost native Atlantic salmon population in the 
western Atlantic. The temperature regimes in these southern rivers made 
possible the tremendous growth and production potential which resulted 
in the historically very large populations in these areas. Historic 
attempts to enhance salmon populations (in GOM rivers) using Canadian-
origin fish failed. This further illustrates the importance of 
conserving native, river-specific populations and the difficulties of 
restoration if they are lost.
    Fay et al. (2006) concluded that this population segment differs 
markedly from other populations of the species in its genetic 
characteristics (Section 6.3.2.3 of Fay et al., 2006). While genetic 
differences were used to examine the ``discreteness'' of this 
population segment, Fay et al. (2006) suggested that the 
``significance'' of these observed genetic differences is that they 
provide evidence of local adaptation. That is, low returns of exogenous 
smolts (i.e., Canadian-origin

[[Page 29348]]

smolts stocked in Maine) and lower survival of smolts from these Maine 
rivers stocked outside their native geographic range (e.g., into the 
Merrimack River) indicate that this population segment is adapted to 
its native environment. Based on this information related to 
significance, Fay et al. (2006) concluded that this population segment 
is significant to the Atlantic salmon species, and therefore, qualifies 
as a DPS (the new GOM DPS) under the provisions of the DPS Policy.
    Fay et al. (2006) (section 6.3.4) explicitly considered whether to 
include hatchery populations in the GOM DPS and concluded that all 
conservation hatchery populations (currently maintained at GLNFH and 
CBNFH) should be included in the GOM DPS. This determination was based 
on the fact that there is a low level of genetic divergence between 
conservation hatchery populations and the rest of the GOM DPS because: 
(1) The river-specific hatchery programs collect wild parr or sea-run 
adults annually (when possible) for inclusion into the broodstock 
programs; (2) broodstocks are used to stock fry and other life stages 
into the river of origin, and, in some instances, hatchery-origin 
individuals represent the primary origin of Atlantic salmon due to low 
adult returns; (3) there is little evidence of introgression from 
Canadian-origin populations; and (4) there is minimal introgression 
from aquaculture fish because of a rigorous genetic screening program 
in the hatchery. Because the level of divergence is minimal, in Section 
6.3.4 Fay et al. (2006) suggested that hatchery populations should be 
considered part of the GOM DPS. However, Fay et al. (2006) also noted 
the dangers of reliance on hatcheries. In short, genetic risks from 
hatcheries include artificial selection, inbreeding depression, and 
outbreeding depression, in addition to other risks such as the 
potential for disease outbreaks, loss of funding, or other catastrophic 
failure at one or more hatcheries. The reader is directed to 
``Population Status of the GOM DPS'' section of this final rule and 
Section 8.5.1 of Fay et al. (2006) for an in depth discussion of these 
risks.
    For the reasons described in Section 6 of Fay et al. (2006), we 
conclude that the GOM DPS as described above warrants delineation as a 
DPS (i.e., it is discrete and significant). Specifically, we conclude 
that the GOM DPS is comprised of all anadromous Atlantic salmon whose 
freshwater range occurs in the watersheds from the Androscoggin River 
northward along the Maine coast to the Dennys River, including all 
associated conservation hatchery populations used to supplement these 
natural populations; currently, such populations are maintained at 
GLNFH and CBNFH. We consider the conservation hatchery populations that 
are maintained at CBNFH and GLNFH essential for recovery of the GOM DPS 
because the hatchery populations contain a high proportion of the 
genetic diversity remaining in the GOM DPS (Bartron et al., 2006). 
Excluded are those salmon raised in commercial hatcheries for 
aquaculture and landlocked salmon because they are genetically 
distinguishable from the GOM DPS. The marine range of the GOM DPS 
extends from the Gulf of Maine to feeding grounds off Greenland. The 
freshwater range of the GOM DPS includes all freshwater bodies in the 
watersheds from the Androscoggin to the Dennys, except those watersheds 
excluded because of natural barrier falls as described in the 
``Delineating Geographic Boundaries'' section of this final rule. The 
most substantial difference between the GOM DPS as listed in 2000 and 
the GOM DPS described in this final rule is the inclusion of the 
majority of the Androscoggin, Kennebec, and Penobscot Basins as well as 
the associated conservation hatchery population at GLNFH.
    Several rivers outside the range of the GOM DPS in Long Island 
Sound and Central New England contain Atlantic salmon (Fay et al., 
2006; section 6.4). The native Atlantic salmon of these areas south of 
the GOM DPS were extirpated in the 1800s (Fay et al., 2006). Efforts to 
restore Atlantic salmon to these areas (e.g., Connecticut, Merrimack, 
and Saco Rivers) involve stocking Atlantic salmon that were originally 
derived from the GOM DPS. Atlantic salmon whose freshwater range occurs 
outside the range of GOM DPS do not interbreed with salmon within the 
GOM DPS, are not considered a part of the GOM DPS, and are not 
protected under the ESA.

Population Status of the GOM DPS

    In evaluating the status of Atlantic salmon, we considered four 
basic attributes that contribute to a viable population: abundance, 
productivity, genetic diversity, and spatial distribution. The 
importance of considering each of these factors is briefly described 
below. However, it is important to note that our ability to conduct 
such analyses for Atlantic salmon is often limited by the availability 
of sufficient data. It is also important to note that the most recent 
data available at the time of writing of this final rule was from 2007. 
We consider the U.S. Atlantic Salmon Assessment Committee (USASAC) 
reports to be the data of record with respect to Atlantic salmon 
counts. USASAC reports are generally not available until several weeks 
after their annual meeting in March. Thus, 2008 data are considered 
only preliminary at the time of writing this final rule.
    Considering abundance levels of a given species is critical to 
evaluating extinction risks. All else being equal, small populations 
are at greater risk of extinction than larger populations because, 
generally, larger populations are better able to withstand the effects 
of environmental variation, genetic processes, demographic 
stochasticity, ecological feedback, and catastrophes (Shaffer, 1981).
    Population growth rate (productivity) provides information 
regarding how a population is performing in the habitat it occupies. In 
evaluating extinction risks, we ideally measure average productivity at 
different life stages and estimates of variance to describe the level 
of uncertainty inherent in the measurements. An example of life stage-
specific data could be smolt emigration estimates which represent: (a) 
The population's potential to increase or (b) the population's ability 
to weather periods of poor marine conditions. Measuring productivity 
rates over time is quite difficult and resource intensive. Therefore, 
simple measures such as spawner population size and replacement rates 
may be used to provide more rapid detection of changes in conditions 
affecting population growth rates.
    For small populations, spatial distribution is important to reduce 
extinction risks from genetic risks and demographic stochasticity. A 
population's spatial distribution depends on habitat quality (including 
accessibility), population dynamics, and dispersal characteristics of 
individuals in the population. Analysis of spatial distribution focuses 
primarily on spawning group distribution (even though spatial 
distribution is important at all life stages) and connectivity of 
populations. Since freshwater habitat is often quite heterogeneous, 
spawning habitat may be distributed as discrete patches. Straying is an 
important component contributing to spatial distribution and, 
typically, straying rates are higher at smaller scales (e.g., occurring 
within subpopulations rather than between populations (Quinn, 1997)).
    Genetic diversity allows species to adapt to a variety of 
environments that provide for the needs of the species and

[[Page 29349]]

protects against short-term environmental change while also providing 
the raw genetic material necessary to survive long-term environmental 
change. Natural demographic and evolutionary processes (patterns of 
mutation, selection, drift, recombination, migration, etc.) are 
important to maintaining a species' genetic diversity.
    The influence of hatcheries on the GOM DPS must be carefully 
considered in evaluating the status of the species. The influence of 
hatcheries can be both positive and negative; we describe these effects 
in some detail below in this section of this final rule. It is 
important, however, to first describe the general operation of 
conservation hatcheries in Maine.
    The USFWS operates two hatcheries in support of Atlantic salmon 
recovery efforts in Maine. Together, Green Lake National Fish Hatchery 
(GLNFH) and Craig Brook National Fish Hatchery (CBNFH) raise and stock 
over 600,000 smolts and 3.5 million fry annually within the range of 
the GOM Atlantic salmon DPS. The primary focus of the conservation 
hatchery program for the GOM Atlantic salmon DPS is to conserve the 
genetic legacy of Atlantic salmon in Maine until habitats can support 
natural, self-sustaining populations (Bartron et al., 2006). As such, a 
great deal of consideration is given to broodstock collection, spawning 
protocols, genetic screening for aquaculture escapees, and other 
considerations as outlined by Bartron et al. (2006). The current 
program started in 1992, when a river-specific broodstock and stocking 
program was implemented for rivers in Maine (Bartron et al., 2006). 
This strategy complies with the North Atlantic Salmon Conservation 
Organization (NASCO) guidelines for stock rebuilding (USASAC, 2005). 
The stocking program was initiated for two reasons: (1) Runs were 
declining in every river in Maine, and numerous studies indicated that 
restocking efforts are more successful when the donor population comes 
from the river to be stocked (Moring et al., 1995); and (2) the numbers 
of returning adult Atlantic salmon to the rivers were very low, and 
artificial propagation had the potential to increase the number of 
juvenile fish in the river through fry and other early life stage 
stocking.
    Current practices of fry, parr, and smolt stocking as well as 
recovery of parr for hatchery rearing are designed to ensure that 
river-specific brood stock is available for future production. Atlantic 
salmon from the Narraguagus, Pleasant, Sheepscot, Machias, East 
Machias, and Dennys populations are maintained at CBNFH in East Orland, 
Maine. These populations are augmented by annual collections of parr 
from their respective natal river; this program is described in detail 
by Bartron et al. (2006). Additionally, returning adult Atlantic salmon 
are trapped at the Veazie Dam on the Penobscot River throughout the 
duration of the run, transferred to CBNFH, and held until spawning in 
the fall of each year. In addition, domestic adults (i.e., offspring of 
the sea-run adults representing all sea-run spawned families) from the 
Penobscot River are maintained at GLNFH in the event that insufficient 
sea-run adults return to the Veazie trap or in the event of a fish loss 
at CBNFH. Adult Atlantic salmon (with the exception of the Penobscot 
River) are maintained in one of six river-specific broodstock rooms at 
CBNFH. Within each broodstock room, adults are maintained separately by 
capture year. Capture year is defined as the year parr were collected 
from a river. Each capture year may represent one to two year classes. 
In addition, fully captive lines, or ``pedigree lines,'' are 
implemented when the recovery of parr from the river environment is 
expected to be too low to ensure future spawning stock is available 
(Bartron et al., 2006). Pedigree lines are established at the time of 
stocking, where a proportional representation of each family from a 
particular river-specific broodstock is retained in the hatchery while 
the rest of the fry are stocked into the river. If parr are recovered 
from the fry stocking for the pedigree lines, individuals are screened 
to determine origin and familial representation and are integrated into 
the pedigree line to maintain some component of natural selection while 
maintaining a broad representation of the genetic diversity observed in 
the broodstock.
    The goals of the captive propagation program include maintenance of 
the unique genetic characteristics of each river-specific broodstock 
and maintenance of genetic diversity within each broodstock (Bartron et 
al., 2006). Evaluation of estimates of genetic diversity within captive 
populations, such as average heterozygosity, relatedness, and allelic 
richness are monitored within the hatchery broodstocks according to the 
CBNFH Broodstock Management Plan (Bartron et al., 2006). Estimates of 
allelic richness within each broodstock have thus far, revealed 
consistent estimates over the brief time series available (generally 
1994 to 2004; Bartron et al., 2007). Information from genetic 
monitoring is used to evaluate management practices to reduce the 
potential for artificially reducing overall genetic diversity. Further 
details of annual genetic monitoring are described by Bartron et al. 
(2007).
    The current low abundance of adult returns, integration of the 
majority of adult returns into the hatchery for the Penobscot, and 
recapture of parr from the wild for broodstock makes the wild and 
hatchery populations interwoven. In the following sections of this 
final rule, we describe the four population attributes of interest 
(abundance, productivity, spatial structure, and genetic diversity) and 
attempt to apply them first to the wild population and then discuss the 
impact the hatchery has on that attribute. For the reasons noted above, 
however, it is rarely possible to completely separate the wild and 
hatchery population in this analysis.

Abundance

    The abundance of Atlantic salmon within the range of the GOM DPS 
has been generally declining since the 1800s (Fay et al., 2006). Data 
sets tracking adult abundance are not available throughout this entire 
time period; however, Fay et al. (2006) in Figure 7.3.1 present a 
comprehensive time series of adult returns to the GOM DPS dating back 
to 1967. It is important to note that contemporary abundance levels of 
Atlantic salmon within the GOM DPS are several orders of magnitude 
lower than historical abundance estimates. For example, Foster and 
Atkins (1869) estimated that roughly 100,000 adult salmon returned to 
the Penobscot River alone before the river was dammed, whereas 
contemporary estimates of abundance for the entire GOM DPS have rarely 
exceeded 5,000 individuals in any given year since 1967 (Fay et al., 
2006).
    Contemporary abundance estimates are informative in considering the 
conservation status of the GOM DPS today. After a period of population 
growth in the 1970s, adult returns of salmon in the GOM DPS have been 
steadily declining since the early 1980s and appear to have stabilized 
at low levels since 2000 (Figure 1). The population growth observed in 
the 1970s is likely attributable to favorable marine survival and 
increases in hatchery capacity, particularly at GLNFH, which was 
constructed in 1974. Marine survival remained relatively high 
throughout the 1980s, and salmon populations in the GOM DPS remained 
relatively stable until the early 1990s when marine survival rates 
decreased, leading to the declining trend in adult abundance observed 
in the early 1990s.

[[Page 29350]]

[GRAPHIC] [TIFF OMITTED] TR19JN09.002

    Adult returns to the GOM DPS have been very low for many years and 
remain extremely low in terms of adult abundance in the wild. Further, 
the majority of all adults return to a single river, the Penobscot, 
which accounted for 91 percent of all adult returns to the GOM DPS in 
2007 (Table 2). As illustrated by Table 3, of the 925 adult returns to 
the Penobscot in 2007, 802 were the result of smolt stocking and only 
the remaining 123 were naturally-reared. The term ``naturally-reared'' 
includes fish originating from natural spawning and hatchery fry 
(USASAC, 2008). Hatchery fry are included because hatchery fry are not 
marked; therefore, they cannot be distinguished from fish produced from 
natural spawning. Because of the extensive amount of fry stocking that 
takes place in an effort to recover the GOM DPS, it is likely that a 
substantial number of fish counted as naturally-reared were actually 
stocked as fry. The term ``hatchery-origin'' includes those fish 
stocked as either parr or smolt from either CBNFH or GLNFH.
    The proportion of naturally reared fish that is attributed to fry 
stocking cannot be determined. Preliminary adult return data for 2008 
(https://www.maine.gov/dmr/searunfish/trapcounts.html) indicated higher 
returns than in previous years, but remain well below conservation 
spawning escapement (CSE) goals that are widely used (e.g., ICES, 2005) 
to describe the status of individual Atlantic salmon populations. When 
CSE goals are met, Atlantic salmon populations are generally self-
sustaining. When CSE goals are not met (i.e., less than 100 percent), 
populations are not reaching full potential, and this can be indicative 
of a population decline. For all rivers in Maine, current Atlantic 
salmon populations (including hatchery contributions) are well below 
CSE levels required to sustain themselves (Fay et al., 2006) (section 
7.1), which is further indication of their poor population status. 
Furthermore, calculation of returns relative to CSE for Atlantic salmon 
include salmon of fry-stocked origin; because these fish are not 
spawned in the wild, displaying returns as a percentage of CSE 
overestimates the degree to which the population is achieving self-
sustainability.

  Table 2--Adult Returns to the Small Coastal Rivers, the Penobscot River, the Kennebec River, and the Androscoggin River From 2001 to 2007. These Data
             are Summarized From Table 3.2.1.2 and Table 16 in the United States Atlantic Salmon Assessment Committee Report (USASAC, 2008)
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                  Small coastal    Penobscot River   Kennebec River     Androscoggin       Total known
                             Year                                    rivers          trap count      trap count \a\   River trap count       returns
--------------------------------------------------------------------------------------------------------------------------------------------------------
2001..........................................................               103               785  ................                 5               893
2002..........................................................                37               780  ................                 2               819
2003..........................................................                76              1112  ................                 3              1191
2004..........................................................                82              1323  ................                11              1416
2005..........................................................                71               985  ................                10              1066
2006..........................................................                79              1044                15                 6              1144

[[Page 29351]]

 
2007..........................................................                53               925                16                20              1014
--------------------------------------------------------------------------------------------------------------------------------------------------------
a Counts not conducted on the Kennebec until 2006.


  Table 3--Adult Returns to Rivers Within the Freshwater Range of the GOM DPS by Origin in 2007. These Data Are
     Summarized From Table 1 in the United States Atlantic Salmon Assessment Committee Report (USASAC, 2008)
----------------------------------------------------------------------------------------------------------------
                           River                             Hatchery-origin  Naturally-reared        Total
----------------------------------------------------------------------------------------------------------------
Androscoggin..............................................                17                 3                20
Kennebec..................................................                 9                 7                16
Dennys....................................................                 2                 1                 3
Narraguagus...............................................                 0                11                11
Other GOM DPS.............................................                 0                39                39
Penobscot.................................................               802               123               925
                                                           -----------------------------------------------------
    Total.................................................               830               184              1014
----------------------------------------------------------------------------------------------------------------

    Declines in both hatchery-origin and naturally reared salmon are 
evident in the Penobscot River (Table 4). Declines in hatchery-origin 
adult returns are less sharp because of the effects of hatcheries. In 
short, hatchery supplementation over this time period has been 
relatively constant, generally fluctuating around 550,000 smolts per 
year (USASAC, 2008). In contrast, the number of naturally-reared smolts 
emigrating each year is likely to decline following poor returns of 
adults. Although it is impossible to distinguish truly wild salmon from 
those stocked as fry, it is likely that some portion of naturally 
reared adults are wild. Thus, wild smolt production would suffer 3 
years after there were low adult returns, because the progeny of adult 
returns typically emigrate 3 years after their parents return. The 
relatively constant inputs from smolt stocking coupled with the 
declining trend of naturally reared adults result in the apparent 
stabilization of hatchery-origin salmon and the decline of naturally 
reared components of the GOM DPS observed over the last 2 decades.

  Table 4--Adult returns, by origin (hatchery-origin and naturally reared) and age (1sw Indicates the Individual Spent One Winter at Sea; 2sw Indicates
   the Individual Spent Two Winters at Sea; 3sw Indicates the Individual Spent Three Winters at Sea; and Repeat Indicates the Individual was a Repeat
                                                    Spawner) to the Penobscot River from 1996 to 2007
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                            Hatchery-origin                        Naturally reared
                             Year                              --------------------------------------------------------------------------------   Total
                                                                   1sw       2sw       3sw     Repeat      1sw       2sw       3sw     Repeat
--------------------------------------------------------------------------------------------------------------------------------------------------------
1996..........................................................       484     1,218         6        18        11       303         3         1     2,044
1997..........................................................       243       934         4        14         4       153         2         1     1,355
1998..........................................................       238       793         0        10        31       133         1         4     1,210
1999..........................................................       223       568         0        11        49       108         0         9       968
2000..........................................................       167       265         0        15        16        69         0         2       534
2001..........................................................       195       466         0         3        21        98         2         0       785
2002..........................................................       363       344         0        15        14        41         1         2       780
2003..........................................................       196       847         1         4         6        56         0         2     1,112
2004..........................................................       276       952        10        16         5        59         3         2     1,323
2005..........................................................       269       678         0         8         6        22         0         2       985
2006..........................................................       338       653         1         4        15        33         0         0     1,044
2007..........................................................       226       575         0         1        35        88         0         0       925
--------------------------------------------------------------------------------------------------------------------------------------------------------

    The influence of CBNFH and GLNFH on abundance of the GOM DPS is 
positive, thus reducing short-term extinction risks to the GOM DPS. 
Below, we briefly describe the three mechanisms by which the 
conservation hatchery programs positively affect the abundance of the 
GOM DPS:
    1. Stocking of large numbers of smolts (Penobscot beginning in 
1974, Dennys beginning in 2001, and Narraguagus beginning in 2008) 
increases adult returns, thus reducing demographic risks (i.e., 
extinction risks) to populations that would otherwise be smaller.
    2. Stocking large numbers of smolts also reduces the risks of 
catastrophic loss because at least one cohort is always at sea and 
could be collected as broodstock in case of a catastrophic event in 
freshwater (e.g., a large contaminant spill) or in a hatchery (e.g., 
disease outbreak).
    3. Rivers without large scale fry stocking efforts have even fewer 
adult returns than those rivers with large scale stocking efforts. 
Further, rivers that lack significant hatchery contributions (fry 
stocking) have not experienced stable

[[Page 29352]]

levels of adult returns since the decline in marine survival in the 
early 1990s. For example, redd counts in the Ducktrap River (a river 
which is not stocked) have been steadily declining since the 1990s to a 
point where no redds were found in the Ducktrap River in 2007, a year 
with favorable conditions for redd counting and over 90 percent of 
spawning habitat surveyed (USASAC, 2008).
    As illustrated by the above data, the abundance of Atlantic salmon 
in the GOM DPS is low and either stable or declining. The proportion of 
fish that are of natural origin is very small (approximately 10 
percent) and is continuing to decline. The conservation hatchery has 
assisted in slowing the decline and helped stabilize populations at low 
levels, but has not contributed to an increase in the overall abundance 
of salmon and has not been able to halt the decline of the naturally-
reared component of the GOM DPS.

Productivity

    The historic productivity of the GOM DPS is unknown. Over long time 
frames, it is expected that productivity fluctuated widely according to 
a diverse range of biotic factors such as food availability and abiotic 
factors such as temperature regime and sea level.
    Contemporary productivity rates for the GOM DPS can be inferred 
from replacement rates. In short, populations with a replacement rate 
of 1.0 or higher are stable or increasing while populations with a 
replacement rate less than 1.0 are declining. The USASAC has estimated 
the replacement rate for the GOM DPS (as listed in 2000) over the last 
several years. Replacement rate for the GOM DPS (as listed in 2000) had 
been below 1.0 for several generations until 2007, when replacement 
rate for the 2002 spawning cohort was 1.47. This translates to on 
average, every adult returning in 2002 replacing itself with 1.47 
adults in 2007. While this increase is promising, it only represents 1 
year; thus, it is premature to conclude that this is indicative of an 
increasing trend.
    Replacement rate is a fairly imprecise measurement of productivity 
for several reasons. First, tracking adult to adult return rates of 
naturally reared fish necessarily includes those fish that result from 
stocking. Thus, it is not true replacement of fish in the wild because 
each river with substantial returns of adults is stocked with fry, or 
smolts as in the case of the Penobscot, Narraguagus, and Dennys Rivers. 
This situation results in an overestimation of productivity (because it 
does not account for the contribution that stocking makes to adult 
returns) and also emphasizes the importance of hatcheries to the 
security of the GOM DPS. Without stocking of hatchery fry and smolts, 
adult returns would presumably be lower and would result in even lower 
replacement rates.
    The influence of hatcheries on productivity is not known with 
certainty, but overall productivity (even with hatchery 
supplementation) is quite low. The first goal of the captive broodstock 
program is to facilitate the recovery of the natural populations and 
minimize the risk of further decline or loss of individual populations 
(Bartron et al., 2006). Over time, more adult returns should 
successfully spawn in the wild and result in replacement rates above 
1.0. However, insufficient data exist to determine whether adult 
returns from hatchery contributions result in more spawners and 
ultimately more truly wild-origin adult returns. The National Research 
Council (NRC, 2004) and the Sustainable Ecosystems Institute (SEI, 
2007) identified this as a key limitation in available data on the 
recovery efforts for salmon in Maine. Without this information, it is 
impossible to estimate, with any certainty, the effect of hatcheries on 
this key population attribute (productivity). Overall, however, 
replacement rates less than 1.0 (as has been the case most years since 
the early 1990s) are indicative of low productivity.
    As illustrated by the above, productivity of the GOM DPS is low and 
has not consistently had a replacement rate above 1.0 such that 
population growth would be expected. There is no current evidence that 
hatcheries have increased or will increase productivity in the wild.

Spatial Distribution

    The historic distribution of Atlantic salmon in Maine has been 
described extensively by Baum (1997) and Beland (1984), among others. 
In short, substantial populations of Atlantic salmon existed in nearly 
every river that was large enough to maintain a spawning population. 
The upstream extent of anadromy extended far into the headwaters of 
even the largest rivers. For example, Atlantic salmon were found 
throughout the West Branch of the Penobscot River as far as Penobscot 
Brook, a distance over 350 river km inland (Atkins, 1870). In the 
Kennebec River, Atlantic salmon ranged as far inland as the Kennebec 
River Gorge and Grand Falls on the Dead River, 235 km inland (Foster 
and Atkins, 1867; Atkins, 1887).
    Today, the spatial structure of Atlantic salmon is limited by 
obstructions to passage and also by low abundance levels. Fish passage 
obstructions caused the decline of many salmon populations (Moring, 
2005). Within the range of the GOM DPS, the Kennebec, Androscoggin, 
Union, and Penobscot Rivers contain dams that severely limit passage of 
salmon to significant amounts of spawning and rearing habitat.
    In addition, the low abundance of salmon within the range of the 
GOM DPS serves to concurrently limit spatial distribution through two 
mechanisms: (1) Lack of sufficient source populations, and (2) hatchery 
limitations. First, in properly functioning salmon populations, some 
areas have relatively abundant salmon populations such that they may 
serve as ``source'' populations. Fish from source populations may seek 
out areas with fewer or no competitors. This is an important dispersal 
mechanism for all anadromous salmonids. Over evolutionary timescales, 
this process led to the colonization of nearly every river in Maine by 
Atlantic salmon. Because the abundance of salmon is so low today, this 
dispersal mechanism is likely not operating and will likely not operate 
until trends in productivity and abundance are reversed. Second, 
spatial distribution is limited today by hatchery capacity. The 
Penobscot River alone would require 12.5 million fry in order to 
properly seed all presently accessible rearing habitat (Trial, 2006), 
while GLNFH and CBNFH can only produce roughly 3.5 million fry annually 
(Barton et al., 2006). Thus, hundreds of thousands of otherwise 
suitable habitat units are currently unoccupied (NMFS, 2008). The 
Sheepscot, Narraguagus, Dennys, Machias, East Machias, and Pleasant 
Rivers are usually stocked with as many fry as are needed to properly 
seed the habitat, although no stocking occurs within a 50-meter buffer 
around areas known to have spawning activity the previous year in order 
to reduce competition between potentially wild and hatchery fry 
(described in detail by Trial, 2006). Hatchery space for the Penobscot 
population is limited by hatchery capacity, such that only 2.5 million 
fry are typically allocated and stocked into the Penobscot River 
annually. Other rivers within the freshwater range of the GOM DPS have 
been stocked to a very limited degree in some years, usually with 
Penobscot-origin fry (see section 5 of Fay et al., 2006, for a detailed 
review).
    The influence of hatcheries on spatial structure of the GOM DPS is 
positive. Without hatchery contributions, fewer juveniles would inhabit 
the rivers of Maine. In section 7.2., Fay et al. (2006)

[[Page 29353]]

examined recent MDMR electrofishing data, which demonstrated that 
rivers with large scale stocking efforts have much higher juvenile 
densities compared to those rivers without large scale stocking 
efforts. The hatchery, therefore, has allowed for ma