Endangered and Threatened Wildlife and Plants; 12-Month Finding on a Petition To List the Cerulean Warbler (Dendroica cerulea) as Threatened With Critical Habitat, 70717-70733 [E6-20530]

Agencies

• Fish and Wildlife Service
• DEPARTMENT OF THE INTERIOR

[Federal Register Volume 71, Number 234 (Wednesday, December 6, 2006)]
[Proposed Rules]
[Pages 70717-70733]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: E6-20530]


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

DEPARTMENT OF THE INTERIOR

Fish and Wildlife Service

50 CFR Part 17


Endangered and Threatened Wildlife and Plants; 12-Month Finding 
on a Petition To List the Cerulean Warbler (Dendroica cerulea) as 
Threatened With Critical Habitat

AGENCY: Fish and Wildlife Service, Interior.

ACTION: Notice of a 12-month petition finding.

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

SUMMARY: We, the U.S. Fish and Wildlife Service (Service), announce a 
12-month finding on a petition to list the cerulean warbler (Dendroica 
cerulea) as threatened under the Endangered Species Act of 1973, as 
amended (Act). The petition also asked that critical habitat be 
designated for the species. After reviewing the best available 
scientific and commercial information, we find that the petitioned 
action is not warranted. We ask the public to submit to us any new 
information that becomes available concerning the status of, or threats 
to, the species. This information will help us monitor and encourage 
the conservation of this species.

DATES: The finding announced in this document was made on November 28, 
2006.

ADDRESSES: Comments and materials received, as well as supporting 
documentation used in the development of this 12-month finding, will be 
available for inspection, by appointment, during normal business hours 
at the Columbia Ecological Services Field Office, 101 Park DeVille 
Drive, Suite A, Columbia, Missouri 65203. Submit new information, 
materials, comments, or questions concerning this species to the 
Service at the above address.

FOR FURTHER INFORMATION CONTACT: Charles Scott, Supervisor (see 
ADDRESSES), by telephone at 573-234-2132, by facsimile at 573-234-2181, 
or by electronic mail at charlie_scott@fws.gov. Individuals who are 
hearing-impaired or speech-impaired may call the Federal Relay Service 
at 1-800-877-8339 for TTY assistance.

SUPPLEMENTARY INFORMATION:

Background

    Section 4(b)(3)(B) of the Endangered Species Act of 1973, as 
amended (16 U.S.C. 1531 et seq.), requires that, for any petition to 
revise the Lists of Endangered and Threatened Wildlife and Plants that 
contains substantial scientific or commercial information that the 
petitioned action may be warranted, we make a finding within 12 months 
of the date of the receipt of the petition on whether the petitioned 
action is: (a) Not warranted, (b) warranted, or (c) warranted, but that 
the immediate proposal of a regulation implementing the petitioned 
action is precluded by other pending proposals to determine whether any 
species is threatened or endangered, and expeditious progress is being 
made to add or remove qualified species from the List of Endangered and 
Threatened Species. Such 12-month findings are to be published promptly 
in the Federal Register. Section 4(b)(3)(C) of the Act requires that a 
petition for which the requested action is found to be warranted but 
precluded shall be treated as though resubmitted on the date of such 
finding, requiring a subsequent finding to be made within 12 months.

Previous Federal Actions

    We added the cerulean warbler to our former Category 2 list of 
candidate species on November 21, 1991 (56 FR 58804). Category 2 
candidate species were those species for which we possessed data 
indicating that proposing to list them as endangered or threatened was 
possibly appropriate, but for which conclusive data on biological 
vulnerability and threat were not available at that time to support 
proposed rules. Category 1 candidate species were those for which we

[[Page 70718]]

possessed sufficient information on biological vulnerability and 
threats to support proposals to list them as endangered or threatened 
species. The cerulean warbler was also in the November 15, 1994, 
Candidate Notice of Review (59 FR 58982) as a Category 2 candidate 
species. The list of Category 2 species was eliminated by the Service 
in 1996. Since then the Service has applied the term ``candidate 
species'' only to those species previously considered to be ``Category 
1'' candidates, and we apply the same definition to these species (61 
FR 7596; February 28, 1996). The cerulean warbler has never been a 
Category 1 candidate species or a candidate species, as defined, since 
1996.
    Due to concerns regarding the population trend of the species, in 
1995, the Service contracted to Dr. Paul Hamel, of the U.S. Forest 
Service's Southern Forest Research Station in Stoneville, Mississippi, 
to develop a cerulean warbler rangewide status assessment report. Dr. 
Hamel completed his report in April of 2000 (Hamel 2000a), and we 
distributed it and posted it on our Web site at that time.
    On November 6, 2000, the Service received an October 30, 2000, 
letter from Douglas A. Ruley of the Southern Environmental Law Center 
in Asheville, North Carolina. Mr. Ruley's letter conveyed a petition to 
list the cerulean warbler as a threatened species and to designate 
critical habitat for the species (Ruley 2000). The following 
organizations were listed as the petitioners: National Audubon Society, 
Defenders of Wildlife, Sierra Club, The Wilderness Society, American 
Lands Alliance, Western North Carolina Alliance, Southern Appalachian 
Biodiversity Project, Appalachian Voices, Cherokee Forest Voices, 
Southern Environmental Law Center, Southern Appalachian Forest 
Coalition, Heartwood, Dogwood Alliance, West Virginia Highlands 
Conservancy, Inc., Virginia Forest Watch, Buckeye Forest Council, 
Allegheny Defense Project, Vernon Civic Association, Conservation 
Action Project, Superior Wilderness Action Network, Indiana Forest 
Alliance, Regional Association of Concerned Environmentalists, Ouachita 
Watch League, Newton County Wildlife Association, Chattooga 
Conservancy, Wild Alabama, Georgia Forest Watch, and South Carolina 
Forest Watch.
    On September 24, 2002, the Service made its initial 90-day finding 
on the petition, and a notice of that finding was published in the 
Federal Register on October 23, 2002 (67 FR 65083). Our finding was 
that the petition presented substantial information indicating that the 
petitioned action of listing the species may be warranted. At that 
time, we initiated a status review, which included a 90-day comment 
period.
    We received 290 responses to our request for additional information 
in our 90-day finding for the cerulean warbler (67 FR 65083; October 
23, 2002). A large number of these responses were identical or similar 
comments. Comments and information were received from 12 State fish and 
wildlife agencies within the range of the warbler, 4 academic 
researchers, 2 county government agencies, the U.S. Forest Service (4 
units), National Park Service (2 units), Department of Defense, U.S. 
Army Corps of Engineers, a U.S. Congressman, 7 corporations, 40 
nongovernmental organizations, numerous private citizens, and several 
other entities. Additionally, we directly contacted, and received 
information from, wildlife agencies and biologists within the cerulean 
warbler's range in Canada and South America. We reviewed all responses 
received, and those that contained new, updated, or additional 
scientific or commercial data were thoroughly considered in this 12-
month finding.
    Due to budget shortfalls during subsequent fiscal years, the 
Service was unable to fund additional work on the petition until late 
in fiscal year 2005. Since that time, we have analyzed the comments 
received after the 2002 finding, reviewed new published and unpublished 
reports and data on the species and factors affecting its habitat, and 
brought together a panel of experts on the species to provide 
additional insight into the current status and trends of the cerulean 
warbler.
    After our resumption of work on the petition in late 2005, a 
lawsuit was filed by five of the petitioners (National Audubon Society, 
Defenders of Wildlife, Southern Appalachian Biodiversity Project, 
Western North Carolina Alliance, and Heartwood) in the U.S. District 
Court for the District of Columbia on February 28, 2006. The suit asked 
the Court, among other things, to compel the Service to make and 
publish in the Federal Register a 12-month finding regarding the 
plaintiffs' petition to list the cerulean warbler as a threatened 
species. Although we had already resumed work on the petition, due to 
the lawsuit, we entered into a settlement agreement with plaintiffs in 
which we agreed to provide our 12-month finding to the Federal Register 
no later than November 30, 2006.

Cerulean Warbler Natural History

    The cerulean warbler is a small insectivorous neotropical migrant 
songbird (11.5 centimeters (4.5 inches) long and weighing 8 to 10 grams 
(0.3 to 0.4 ounces)). It breeds in mature deciduous forests primarily 
within the central hardwood region of eastern North America, primarily 
in the Ohio and Mississippi River Valleys and adjacent areas east of 
the Appalachians, in New England and southern Canada, and in the Great 
Lakes region. (Hamel 2000a, pp. 2-4). The breeding range generally 
extends from the eastern Great Plains, north to Minnesota, east to 
Massachusetts, and south to North Carolina and Louisiana (Hamel 2000a, 
p. 2), encompassing 33 States and 2 Canadian Provinces. The core area 
of the breeding range is currently within the Cumberland Plateau and 
Ohio Hills physiographic regions in eastern Tennessee, eastern 
Kentucky, southern and western West Virginia, southeastern Ohio, and 
southwestern Pennsylvania (Villard and Mauer 1996, p. 7 and Figure 7; 
Sauer et al. 2005a). This species undertakes a long migration compared 
to many other warblers and passerines of similar size (Hamel 2000b, p. 
1), covering a distance of approximately 4,000 kilometers (km) (2,500 
miles (mi)) between the central latitudes of North America and northern 
latitudes of South America. The migratory pathway between the breeding 
and wintering grounds is not well known, but for most individuals, it 
likely includes a flight across the Gulf of Mexico and stops at a 
limited number of locations in Central America and northern Colombia or 
Venezuela (Hamel 2000b, p. 4). The fall migration to South America 
might be along a more easterly path than that of the northward 
migration in the spring (Dunn and Garrett 1997, p. 405). Cerulean 
warblers winter in broad-leaved evergreen forests within a relatively 
narrow band of middle elevations (500 to 1,800 meters (m); 1,650 to 
5,900 feet (ft)) in the northern Andes Mountains in Venezuela, 
Colombia, Ecuador, Peru, and Bolivia and possibly in the Guayana 
Highlands of southeastern Venezuela, especially the tabletop mountains 
(tepuis) of this ecoregion (Robbins et al. 1992, p. 559; Moreno et al. 
2006 unpublished report, p. 3).
    On the breeding grounds, cerulean warblers prefer mature hardwood 
forests with tall, large-diameter trees and a structurally diverse 
canopy (multiple vegetation layers, often associated with uneven-aged 
forest stands). They occupy forests with these structural 
characteristics in both upland and

[[Page 70719]]

bottomland locations (Hamel 2000b, p. 4). In the Appalachian Mountains, 
they tend to occur more frequently and in higher abundance on ridge 
tops than in valley bottoms (Weakland and Wood 2005, pp. 503-504; Wood 
et al. 2006, pp. 160-161; Buehler et al. in press, p. 9). Throughout 
much of their breeding range, they prefer to breed in large forest 
patches, and so are considered ``area-sensitive'' (Robbins et al. 
1989a, p. 25; Mueller et al. 2000, p. 15), although they might not be 
as sensitive to forest patch size in well-forested and less fragmented 
landscapes where avian nest predation and parasitism rates tend to be 
lower (Hamel 2000b, p. 4). In parts of their range, cerulean warblers 
exhibit positive associations with canopy gaps and relatively small 
internal forest openings (Perkins 2006, p. 26), but they avoid abrupt 
edges between forest and large areas of open land (Wood et al. 2006, p. 
160). Post-fledging habitat for this species has not been studied, but 
assuming cerulean warblers are similar to other mature forest-
associated birds, they might seek out areas where shrubby vegetation 
provides good cover from predators as well as an abundance of good 
foraging substrate. Such areas might include small forest openings or 
early successional habitats, but habitat use during this period of the 
year has not been described and the relative importance of different 
habitat types during the post-fledging period is not known.
    Insects are the primary food source of cerulean warblers throughout 
the year. During the breeding season, their diet has been observed to 
consist primarily of Homoptera and Lepidoptera but also may include 
small amounts of Coleoptera, Hymenoptera, Diptera, Hemiptera, Araneae, 
and other arthropods (Hamel 2000b, p. 6). While no detailed studies of 
diet have been completed during the non-breeding period, cerulean 
warblers appear to use nectar resources, as well as insects, during at 
least some period of their residency on their non-breeding grounds in 
South America (Jones et al. 2000, p. 961; USFWS 2006, Appendix 5--M.I. 
Moreno's PowerPoint presentation, slide 15) and have also been observed 
eating small amounts of plant material during migration (Hamel 2000b, 
p. 5). Their primary foraging mode for capturing insects is gleaning 
prey from the upper and lower surfaces of leaves. They also use 
sallying and hover-gleaning to a lesser extent (Hamel 2000b, p. 5).
    Cerulean warblers build their nests high above ground (mean height 
of 11.4 m (37 ft); Hamel 2000b, p. 9) in the mid-story or canopy of 
trees. Clutch size is normally 3 or 4 eggs with an incubation period of 
11 to 12 days and a nestling period of 10 to 11 days. Their nests are 
known to be parasitized by brown-headed cowbirds, particularly in the 
western portion of the cerulean warbler breeding range where cowbirds 
are more abundant (Hamel 2000b, pp. 9-11). Nest success varies annually 
and regionally, with observed average annual nest success rates at 
specific study sites ranging from approximately 20 percent in southern 
Indiana and the lower Mississippi River valley to approximately 58 
percent in Ontario and eastern Tennessee. The average number of young 
fledged per successful nest also varies, although somewhat less 
dramatically, with reports of annual values between 1.7 and 3.0 for 
most study sites (USFWS 2006, Appendix 5--D. Buehler's PowerPoint 
presentation, slides 25-28). Cerulean warblers typically arrive on 
their breeding grounds between mid-April and mid-May, depending on 
latitude, and remain there until sometime between late July and mid-
September (Dunn and Garrett 1997, pp. 405-406). Cerulean warblers 
usually raise a single brood during this period; multiple nesting 
attempts are commonly undertaken if initial nest attempts fail. It is 
rare for this species to raise two broods in the same breeding season.
    Cerulean warblers are predominantly socially monogamous (one male 
mated with one female), but social bigamy (one male mated with two 
females) has been observed in the Ontario population (USFWS 2006, 
Appendix 4, Day 2-p. 2). This behavior has not been studied at other 
locations. Some researchers have also observed a clumped distribution 
of cerulean warbler territories within study sites, apparently 
independent of habitat features. However, these patterns have not been 
studied rigorously nor confirmed as being different from a random 
distribution or a result of habitat selection (Hamel 2000b, p. 8).
    Analysis of genetic variability at the population level has 
revealed no significant variation in neutral genetic markers across the 
breeding range, suggesting a single genetic population for this species 
(Veit et al. 2005, pp. 165-166). A study of natal and breeding 
dispersal between years using stable isotope analysis corroborates this 
hypothesis by suggesting a relatively high level of interannual adult 
dispersal between regions, particularly within the central portions of 
the breeding range (USFWS 2006, Appendix 4, Day 1--p. 14). Adult 
dispersal to different breeding locations between years appears to be 
lower in both the southern and northern portions of the range than in 
the center of the range, suggesting higher site fidelity to breeding 
locations in those portions of the range. Natal dispersal between 
regions within the breeding range did not appear to be any more 
pronounced than adult dispersal. This is different than many other 
warbler species, which typically exhibit much higher natal dispersal 
than adult dispersal. Dispersal characteristics of cerulean warblers 
probably influence source-sink dynamics of the population, and more 
information on dispersal is needed to understand the current population 
trend of the species.
    On the wintering grounds, this species may prefer forests with old-
growth conditions, but it has also been found in second-growth forests 
and shade-grown coffee plantations (Hamel 2000b, p. 5; Jones et al. 
2000, p. 958). As with its breeding habitat, a structurally diverse 
canopy with multiple vegetation layers appears to be an important 
component of its wintering habitat. It is generally found in mixed-
species flocks of canopy-dwelling birds, and this association with 
mixed-species flocks could be an important characteristic of their 
occurrence on the wintering grounds (Hamel 2000b, p. 5), although more 
study of their social behavior is needed. Cerulean warblers usually 
reside on their winter grounds from October to February (Hamel 2000b, 
p. 9--Figure 3).
    Cerulean warblers are nocturnal migrants. Little is known about 
habitat preferences and other ecological aspects of this bird's 
migration. Several stop-over locations for spring migration have been 
found in Belize (Parker 1994, p. 70), Honduras, and Guatemala (Welton 
et al. 2005, p. 1), but records of this species during migration 
elsewhere are scarce. To explain this, one hypothesis is that cerulean 
warblers could migrate in pulses of large groups of individuals that 
make relatively long flights between stops (for example, northern South 
America to middle Central America and then across the Gulf of Mexico to 
southern United States). Even fewer records exist for cerulean warblers 
during the southward migration in the fall, prompting the suggestion 
that these birds might fly non-stop from the southern U.S. all the way 
to the northern coast of South America. Isotope analyses indicate some 
level of migratory connectivity for this species (USFWS 2006, Appendix 
4, Day 2--pp. 7-8), suggesting that individuals residing in the 
northern portions of the breeding range tend to go to more northerly 
portions of the wintering range and birds from the southern portions of 
the breeding range go to the

[[Page 70720]]

more southerly portions of the wintering range.
    Survival rates of cerulean warblers have not been studied widely 
across their range. Only one study has published estimates of minimum 
survival rates. Jones et al. (2004, p. 17) reported an annual adult 
male survival rate of 0.49 over the period 1995 to 2001; or 0.54 in 
``normal years'' and 0.40 following an ice storm in 1998. These 
estimates are minimum values because they do not account for adult 
dispersal and emigration between breeding seasons.

Population Size and Trends

Background
    Since its inception in 1966, the North American Breeding Bird 
Survey (BBS) is the primary data source for estimating population 
trends of more than 400 species of birds breeding in North America 
(Droege 1990, p. 1). More than 4,000 BBS survey routes are distributed 
along secondary roads across the United States and southern Canada in a 
stratified random design. Each year, volunteer observers count birds 
along these routes, following standardized protocols. Surveys are 
conducted at approximately the same time each year, which is typically 
during the first half of June in most locations. Each survey route 
consists of 50 stops spaced 0.8 km (0.5 mi) apart. Observers count all 
the birds seen and heard within 0.4 km (0.25 mi) of each stop location 
during a three-minute period (Droege 1990, p. 1). The sum of the counts 
for each species over the 50 stops is used as an index of relative 
abundance for that route (Link and Sauer 2002, pp 2833).
    Statistical analyses are performed on these index values across 
routes to estimate population trends for particular species or groups 
of species. Two statistical analysis techniques are currently employed 
by analysts working with the BBS data: The route-regression method 
(Geissler and Sauer 1990, pp. 54-56) and the hierarchical model method 
(Link and Sauer 2002, pp. 2,833-2,836). The hierarchical model method 
is the more recently developed method, and BBS analysts are in 
transition from using the route-regression method to using primarily 
the hierarchical model method, which is a less subjective and more 
efficient method for estimating trend (Link and Sauer 2002, p. 2,837). 
The presentation of BBS data in the 2000 petition (Ruley 2000) used the 
route-regression method. Throughout this finding we discuss BBS data 
using the newer hierarchical model method. As a result, the figures 
used herein to describe BBS population trends differ from those used in 
the petition. Statistical analyses can be conducted across different 
time frames and spatial scales (for example, States, bird conservation 
regions, range-wide).
    It is important to recognize that the BBS was designed to estimate 
trends (changes in population) and not actual abundance (population 
size) of birds. Much of the criticism that has been leveled at the 
BBS--including doubts expressed about the BBS in the Service's positive 
90-day finding on the petition to list the cerulean warbler--stems from 
confusion about the survey's objective and the protocols required to 
meet that objective. The following discussion addresses four aspects of 
the BBS that contribute to this confusion and why these issues do not 
detract from the usefulness of BBS for tracking bird population trends.
    (1) The point count survey methodology of the BBS does not result 
in a complete count of the birds present. The efficiency with which 
birds are counted varies between observers and within observers over 
time and space. In addition, a 3-minute count is not long enough to 
detect all birds present in a given location due to temporal 
variability (both daily and seasonally) in detectability of different 
species. However, the BBS methodology does provide an index of relative 
abundance of birds along the survey routes. This index can be scaled to 
different levels of abundance using different analysis methods and 
provides an appropriate means for assessing population change along the 
routes. An index of relative abundance is suitable for tracking changes 
in the size of the entire population if the ratio between the number of 
birds detected in the surveys and number of birds actually present 
across the landscape remains fairly constant and without any 
directional bias across years (Bart et al. 1998, pp. 212-214).
    The statistical analyses of BBS data help to address some of the 
limitations pertaining to observer efficiency by incorporating 
variables that account for observer effects into the analyses. Such 
effects as differences in counts between observers in different years 
on the same route or the differences between an observer's first count 
and counts in subsequent years on the same route (the novice effect) 
are accounted for in the statistical analysis of the survey data (Sauer 
et al. 1994, pp. 59-60; Link and Sauer 2002, p. 2,834).
    Another factor contributing to incomplete counts of all the birds 
present is that most detections of forest-associated songbirds are 
largely through observers hearing the songs of males. Females of most 
forest songbirds do not sing and, therefore, are more difficult to 
detect during the breeding season. Thus, females of these species are 
greatly undersampled by the BBS. Again, this limitation is not relevant 
to the detection of population trends as long as trends in the male 
portion of the population are representative of trends in the entire 
population. For most small songbirds, such as the cerulean warbler, 
there is no substantial data indicating either a highly skewed sex 
ratio or a large difference in survival rates between the sexes such 
that trend data might be biased.
    (2) BBS surveys are conducted along roadsides and might not 
accurately reflect habitats across entire landscapes. The proportion of 
different habitat types could be different across landscapes compared 
with what is sampled by BBS routes. However, this limitation, in and of 
itself, does not render the BBS ineffective in estimating trends of 
forest birds unless there is a consistent bias in the rate of change of 
habitats bordering roads compared to change of habitats away from 
roadsides. The fact that birds that avoid habitat edges might not be as 
abundant near roads as away from roads also does not influence trend 
estimates, except perhaps to reduce overall sample size for such 
species and require more years of data or more detections to achieve 
appropriate levels of statistical significance.
    Experimental studies comparing roadside with off-road counts or 
modeling efforts to assess relative amounts of different habitats in 
the areas immediately surrounding BBS survey routes and areas away from 
routes are necessary to address the issue of roadside habitat bias for 
the BBS. Two published studies have evaluated the bias associated with 
roadsides in the eastern United States. These studies were conducted in 
Ohio and Maryland. They both concluded that, although BBS routes under-
sampled forest habitats in the regions evaluated (areas adjacent to BBS 
routes tended to have proportionately less forest cover than did the 
region as a whole), they did not find a bias in the change in habitats 
over time along BBS roadside routes compared with the larger landscapes 
surrounding those routes (Bart et al. 1995, p. 760; Keller and Scallan 
1999, pp. 53-55). These studies suggest that the roadside nature of the 
BBS does not create a substantial bias in the BBS data pertaining to 
habitat changes that are likely to influence bird population trends. In 
contrast with this apparent lack of bias in trend estimates, the 
indication from these studies that BBS routes might under-sample forest

[[Page 70721]]

habitats in the East could have implications for the population size 
estimates based on the Partners in Flight method (discussed below). 
However, an unpublished study from West Virginia (Weakland et al. 2003, 
p. 8) found no significant difference between the abundance estimates 
of cerulean warblers from off-road counts and from BBS routes. The 
study found a tendency for the off-road counts to be higher than counts 
on BBS routes, but the difference was not significant. The study 
concluded that, for cerulean warblers, data collected on BBS routes in 
West Virginia are comparable to data collected from off-road locations 
(Weakland et al. 2003, p. 8).
    In the positive 90-day finding on the petition to list the cerulean 
warbler, the Service expressed doubt on the ability of BBS data to 
reliably determine bird population trends of mature forest-associated 
species, such as the cerulean warbler. Reasons for this doubt were 
primarily associated with concerns about a possible roadside bias and 
concerns about lack of uniform coverage of BBS routes across the range 
of the cerulean warbler. To date, the published evidence on the topic 
of the roadside bias suggests that the roadside nature of the BBS does 
not significantly bias its ability to accurately track population 
trends of mature forest species, such as cerulean warblers (Bart et al. 
1995, p. 760; Keller and Scallan 1999, pp. 53-55). Furthermore, the 
more recently implemented hierarchical model method for analyzing BBS 
data estimates trends more efficiently (resulting in smaller confidence 
intervals around the trend estimate) based on the available data (Link 
and Sauer 2002, p. 2837), reducing concerns about lack of uniformity in 
coverage of BBS routes, particularly at the rangewide scale.
    It is also worth noting that efforts to compare population trends 
calculated from BBS data with independent data sources have 
corroborated the trends indicated by the BBS for a variety of species, 
including independent trends based on the Christmas Bird Count, 
Mourning Dove Survey, raptor migration counts, and checklist programs 
(Droege 1990, p. 3). In addition, many peer-reviewed publications have 
been completed using BBS data (for example, Robbins et al. 1989b, Sauer 
et al. 1994, Link and Sauer 1997, Link and Sauer 1998, Royale et al. 
2002, Sauer and Link 2002), indicating the overall robustness and 
scientific credibility of the BBS and its utility for monitoring bird 
population trends.
    (3) A published analysis of BBS data using the hierarchical model 
method indicates that at the range-wide level, cerulean warblers have 
declined at an average rate of 3.04 percent per year during the period 
of 1966 to 2000, with the 95 percent credible interval (confidence 
interval for hierarchical method; C.I.) for the trend estimate being -
4.02 to -2.07 (Link and Sauer 2002, p. 2837). A more recent, but 
unpublished, analysis of the BBS data for the years 1966 to 2005 using 
the hierarchical model method indicates a similar result: cerulean 
warbler trend was -3.2 percent per year (95 percent C.I.: -4.2 to -2.0) 
for this 40-year period (USFWS 2006, Appendix 5, slide 21 of J. Sauer's 
PowerPoint presentation). This recent estimate was based on data from 
243 BBS routes on which cerulean warblers were detected at least once 
during that 40-year period. The rangewide relative abundance reported 
from this recent analysis was 0.25 birds per route, which is relatively 
low (less than 1 bird per route), and warrants some caution when 
considering the BBS results for this species, because a positive bias 
in the trend might occur with low counts, and because the variances are 
imprecise (Sauer et al. 2005b). Within the core of the species' range 
in the Appalachian Mountains (Bird Conservation Region 28), which 
currently supports an estimated 80 percent of the breeding population 
(as calculated using the methods described by Rosenberg and Blancher 
2005), the relative abundance from the recent analysis was 1.03 birds 
per route and the 40-year trend was -3.1 percent per year (95 percent 
C.I.: -4.4 to -1.7; USFWS 2006, Appendix 5, slides 17-19 from J. 
Sauer's presentation).
    Analysis of the rangewide trend over the last 10 years (1996 to 
2005) compared with the previous 30 years (1966 to 1995) indicated no 
significant change in the trend between those two periods (estimated 
change in trend = -0.5 percent, 95 percent confidence interval = -3.8, 
+3.4). The trend estimate for cerulean warblers over the first 30 years 
of the BBS was -3.0 percent per year (C.I.: -4.3, -1.8) and the 
estimate for the past ten years was -3.6 percent per year (C.I.: -6.3, 
-0.1). Because 10 years is a smaller sample size than 30 years, the 
trend estimate based on the last 10 years is less precise than the 
estimate from the previous 30 years, so that the 10-year credible 
interval completely overlaps the 30-year credible interval. Thus, the 
available data suggest that the trend for cerulean warblers has not 
changed during the more recent period and the population continues to 
decline by about 3 percent per year, including within the Appalachian 
core region (Sauer 2006).
    (4) Partners in Flight produced estimates of global population size 
for North American land birds (Rich et al. 2004, pp. 69-77) based on a 
method developed by Rosenberg and Blancher (2005, pp. 58-61). The 
estimate of the cerulean warbler population was 560,000 individuals 
based on an average of counts made on BBS routes during the period of 
1990 to 1999; it can be thought of as an estimate for the year 1995 
(the mid-point of the time period). Partners in Flight rated the 
relative accuracy of their population estimates based on known sources 
of variation and limitations of the methodology pertaining to each 
species. Statistically derived confidence limits could not be provided 
because the variance has not been measured for some of the parameters 
and assumptions used in the method. Partners in Flight rated the 
accuracy of the population estimate for cerulean warblers as 
``moderate,'' suggesting that they felt the estimate was likely to be 
within the correct order of magnitude (100,000's of birds rather than 
millions or 10,000's of birds) and could be within 50 percent of the 
true number (for example, 280,000 to 840,000).
    The Partners in Flight method uses BBS relative abundance data 
along with several assumptions and correction factors to calculate the 
estimated population size for species covered by the BBS (Rosenberg and 
Blancher 2005, pp. 58-61). The method is based on the idea that, at 
each stop on a BBS route, an observer is recording birds within 400m 
(1,300 ft ) of that stop location (per BBS survey protocol). Thus, the 
observer is effectively sampling an area equal to a circle with a 400m 
(1,300 ft) radius. Over the 50 stops of a BBS route, this sums to an 
effective sampling area of 25.1 km\2\ (9.7 mi\2\). After making some 
assumptions regarding BBS routes adequately representing habitats 
across large landscapes and assumptions about the detectability of 
birds, the average number of birds counted on BBS routes within a 
particular region can be extrapolated across that region to calculate 
an estimated population size.
    The following paragraphs present a list of the primary assumptions 
of the Partners in Flight method and discussion of the effects 
violations of these assumptions are likely to have on calculations of 
cerulean warbler population estimates.
    (a) BBS routes are distributed randomly across regional strata. The 
BBS methodology prescribes random distribution of survey routes within

[[Page 70722]]

sampling strata, and the assumption that BBS routes are randomly 
distributed has not been questioned. However, the intensity of route 
allocation within particular strata and the topographic location of 
routes are two factors that could lead to biased population estimates. 
For example, if BBS routes in the Appalachian Mountains tend to be 
along roads that follow creek bottoms, and if cerulean warblers tend to 
be more abundant on ridge tops, as indicated in Weakland and Wood 
(2005, pp. 503-504), Wood et al. (2006, pp. 160-161), and Buehler et 
al. (in press, p. 9), then the BBS counts could be biased by 
undersampling the topographic locations where these birds are likely to 
be most abundant. Both the route allocation and topographic location 
biases could lead to an underestimate of total cerulean warbler 
population size.
    (b) BBS routes sample habitats in proportion to their relative 
amounts within the regional strata. The possibility of a habitat bias 
from the roadside nature of BBS routes contributes to uncertainty about 
the accuracy of population estimates derived from the Partners in 
Flight method. As discussed above in relation to population trend 
estimation, the two studies that have been conducted in the eastern 
United States have shown that BBS routes in Ohio and Maryland 
undersample forest habitats compared to the surrounding landscape (Bart 
et al. 1995, pp. 759-761; Keller and Scallan 1999, pp. 53-55). If a 
similar bias toward underrepresenting forest habitat exists throughout 
much of the cerulean warbler's range, then such a bias would result in 
an underestimation of the total population size when using the Partners 
in Flight method. Various efforts are underway to evaluate the habitat 
bias of BBS routes across much of the United States, but results are 
not available yet.
    (c) Detectability of different bird species is a function of their 
distance from the observer and time of day, and all species have a 
fixed, average maximum detection distance. Correction factors for 
detection distance and time of day were incorporated into the 
estimation method to address this assumption. For the detection 
distance, species were assigned to one of five categories corresponding 
to different average maximum distances at which these birds were likely 
to be detected based on habitat type, song quality, and likelihood of 
being detected in some way other than by song (for example, hawks 
soaring in the distance): 80m (260 ft), 125m (400 ft), 200m (650 ft), 
400m (1,300 ft), and 800m (2,600 ft). These different detection 
distances result in different effective sampling areas for BBS routes. 
Cerulean warblers were assigned a detection distance of 125m (400 ft), 
which is the assumed average maximum distance at which an observer will 
be able to detect a singing bird. This assumption has not been tested, 
and some experts believe that this detection distance might be an 
overestimate of the distance at which a singing cerulean warbler can 
always be heard; it is unlikely to be an under-estimate (USFWS 2006, 
Appendix 4, Day 2--pp. 1-2). If the real maximum detection distance for 
this species is less than 125m (400 ft), it would result in a larger 
population estimate based on the Partners in Flight method. For 
example, using a detection distance of 100m (325 ft) would result in a 
population estimate that is approximately 60 percent higher than the 
estimate using a 125m (400 ft) detection distance. The large influence 
of relatively small changes in detection distance on the resulting 
population estimate indicates that detection distance is a critical 
parameter in the population estimation methodology and contributes a 
large amount of uncertainty pertaining to the population estimate for a 
particular species when the accuracy of this parameter is unknown.
    To correct for detection issues associated with time of day, 
Rosenberg and Blancher (2005, pp. 59-61) developed distribution curves 
of the detections for each species over the 50 stops of BBS routes. 
Based on these curves, peak detection probabilities were determined for 
each species and then a ratio of the peak detections to average 
detections was calculated. This ratio is used to adjust the average 
numbers of birds detected per route to peak numbers per route, 
reflecting numbers that would be expected if the peak detection 
probability lasted throughout the morning hours when BBS routes are 
surveyed. The time of day correction factor calculated for cerulean 
warbler is 1.35 (Rosenberg and Blancher 2005, p. 63--Table 2). The 
methods for deriving this correction factor are empirically based, and 
there is little reason to believe that it is biased or otherwise 
inappropriate for cerulean warblers.
    One potential correction factor that was not incorporated into the 
Rosenberg and Blancher (2005) method and that could influence 
population estimates for cerulean warblers is a correction for 
detectability associated with the season. The song rate of most 
cerulean warbler males declines once they become mated and as the 
breeding season progresses (USFWS 2006, Appendix 4, Day 2--p. 2). The 
breeding season typically begins between mid-April and early May 
throughout much of the breeding range. Most BBS routes are run during 
the first half of June, and overall song rate of mated males is likely 
to be lower at that time than earlier in the breeding season. Such a 
time of season effect could contribute to an under-estimate of the 
total cerulean warbler population size.
    (d) Individuals detected during a count represent one member of a 
pair. A pair correction factor of two times the initial estimate was 
also incorporated into the method to address Assumption D. Most 
individuals in breeding populations are mated during the time of the 
BBS survey, but it is usually only one member of each pair that is 
detected (for example, a singing male). Rosenberg and Blancher (2005, 
p. 61) acknowledge that the appropriate pair correction factor for all 
species is somewhere between one and two, because not all individuals 
in a breeding population are mated. However, this correction factor has 
not been empirically established for any species yet. Field studies 
indicate that not all male cerulean warblers attract mates during the 
breeding season, although some males of this species are also known to 
be bigamous (USFWS 2006, Appendix 4, Day 2--p. 2). The proportion of 
unmated and bigamous males across the species range is unknown. The 
most appropriate pair correction factor for cerulean warblers might be 
a number less than two, but insufficient data currently exist to 
estimate what this number should be for the entire population. A pair 
correction factor less than two would result in a smaller population 
estimate, while a pair correction factor greater than two would result 
in a larger population estimate.

Status of the Cerulean Warbler Population

    We used a stepwise approach to evaluate what single factor or 
combination of factors may affect the cerulean warbler's population 
trend in order to evaluate whether the species warrants listing as 
threatened or endangered under the Endangered Species Act. First, we 
used all available information, including that contained within the 
petition, scientific literature, and expert opinion (USFWS 2006) to 
identify potential factors that might explain the historical and 
projected population trends (see previous section ``Population Size and 
Trend''). Next, we gathered information to assess whether the 
likelihood of occurrence or magnitude of effect of the factors were 
likely to result in population-level effects. We used the qualitative 
judgments of independent experts (USFWS 2006) to assess these potential

[[Page 70723]]

causal factors where quantitative data are unavailable. Then, we 
synthesized the information on the past and future factors with 
estimates of historical (Link and Sauer 2002, p. 2837, Sauer 2006) and 
projected (Thogmartin 2006) cerulean warbler population trends to 
estimate to what degree potential factors might influence the species' 
risk of extinction. Finally, we compared the results of our analysis to 
the five factors listed in the Act to ensure thorough consideration of 
potential threats, and, in light of the Act's five-factor analysis, we 
evaluated whether the species' current or projected status met the 
definitions of threatened and endangered.

Summary of Factors Affecting the Species

    Section 4 of the Act (16 U.S.C. 1533) and our implementing 
regulations at 50 CFR part 424 set forth the procedures for adding 
species to the Federal endangered and threatened species list. A 
species may be determined to be an endangered or threatened species due 
to one or more of the five factors described in section 4(a)(1), as 
follows: (A) The present or threatened destruction, modification, or 
curtailment of its habitat or range; (B) overutilization for 
commercial, recreational, scientific, or educational purposes; (C) 
disease or predation; (D) the inadequacy of existing regulatory 
mechanisms; or (E) other natural or manmade factors affecting its 
continued existence. In making this finding, information regarding the 
status of, and threats to, the cerulean warbler in relation to the five 
factors is discussed below.
    In developing our 12-month finding for the cerulean warbler, we 
considered all scientific and commercial information on the status of 
the species that we received during the comment period following our 
90-day finding. We also searched the scientific literature for relevant 
data and consulted experts on the cerulean warbler and threats to its 
habitat to ensure that this finding is based on the best scientific and 
commercial data available.
    As noted earlier, we considered the population trend estimate of -
3.2 percent per year (CI = -4.2 to -2.0), which is based on Breeding 
Bird Survey data (Link and Sauer 2002, p. 2837; Sauer unpublished data 
2006), to be the best available representation of the species 
population status. This trend estimate comprises all of the factors 
causing population change during the 40-year period of Breeding Bird 
Survey data collection. In other words, all the factors affecting 
cerulean warbler demographics have combined over the past 40 years to 
produce an annual average decline of 3.2 percent per year, with 90 
percent certainty that the true decline is between 4.2 and 2.0 percent 
per year (Link and Sauer 2002, p. 2837; Sauer unpublished data 2006). 
The information available suggests that the factors described in this 
section will continue affecting cerulean warbler habitats and 
demography in a similar manner, resulting in a continuing population 
decline of approximately 2 to 4 percent per year.
    We describe the potential contributing factors to the species' 
approximately 3 percent annual decline in the following description of 
the five listing factors (iterated above).

A. The Present or Threatened Destruction, Modification, or Curtailment 
of Its Habitat or Range

    After consideration of all available information, the Service has 
determined that four biological mechanisms operating throughout the 
species' annual range are likely to be primary contributors to the 
species' declining population trend. Each of these mechanisms is 
related to changes in habitat in North America, South America, and 
along the species' migration routes. These mechanisms are:
    1. Reduction in available nesting sites and suitable breeding 
territory characteristics because of loss or degradation of habitat,
    2. Reduction in foraging success resulting from decreased prey 
abundance, primarily on the wintering ground in South America,
    3. Increased predation throughout the species' annual range and 
nest parasitism of cerulean warblers in the breeding grounds, resulting 
from habitat fragmentation, and
    4. Loss of migration habitat.
    Each of these four mechanisms results, either directly or 
indirectly, from the reductions in quality and quantity of cerulean 
warbler habitat (Factor A of the Act) and therefore, all will be 
discussed under Factor A.
    1. Reduction in available nesting sites and suitable breeding 
territory characteristics because of loss or degradation of habitat:
    Although we do not have a rangewide numerical relationship between 
habitat loss and population change, we do know that there is a positive 
relationship between cerulean warbler nest presence and mature and old-
growth hardwood forests with large trees, small gaps, and vertical 
diversity in vegetation layers (Hamel 2000b, pp. 12-18; Weakland and 
Wood 2002, p. 13). Therefore, we can conclude that degradation or 
removal of suitable mature and old-growth hardwood forestland will 
result in reductions in nesting opportunities, and that accumulation of 
habitat losses is likely to result in declines in cerulean warblers.
    We do not know what happens to individual birds when breeding 
habitat is removed. Displacement of adults and mortality of nestlings 
is likely if removal of nesting stands occurs during the breeding 
season. Nestling or post-fledging mortality may also occur if habitat 
within nesting territories is eliminated or quality is reduced below an 
unknown threshold level. Results of recent studies suggest that 
cerulean warblers are capable of interannual movement (Veit et al. 
2005, pp. 165-166; USFWS 2006, Appendix 5f, slide 17 of Jones 
PowerPoint); therefore, breeding habitat loss during the non-breeding 
season is likely to result in relocation of adults that return during 
the subsequent breeding season. However, the degree to which 
reproductive success or survival of displaced individuals is affected 
is likely dependent upon several variables, including whether the 
displaced birds relocate into already occupied or unoccupied, or 
whether remaining habitat is optimal or suboptimal. We do not have 
information to assess the degree and type of impact of breeding habitat 
of site-specific habitat loss, unless known occupied nests are removed.
    Degradation of habitat quality can occur at several scales, and the 
resulting effect on cerulean warblers is likely to be context-
dependent. Loss of a single dominant tree in a stand possessing 
numerous other dominant trees may have little or no effect on the 
reproductive success of breeding cerulean warblers, whereas loss of a 
single dominant tree in a stand having few other large trees may render 
a formerly suitable site unsuitable for nesting birds. Context is 
probably similarly important at larger scales. Reduction in patch size 
and introduction of hard edges may result in greater local population 
declines and habitat unsuitability where a forest stand is surrounded 
by an already fragmented landscape as opposed to largely intact forest. 
Thus, habitat content factors that operate at local scales (to include 
nest trees, prey base, etc.) and habitat context factors that operate 
at larger scales (to include things like habitat patch size, degree of 
landscape fragmentation, etc.) are both important determinants of 
overall habitat quality for breeding cerulean warblers.

[[Page 70724]]

    The amount, distribution, and quality of habitat for breeding 
cerulean warblers has been altered dramatically since European 
settlement in the early 1600s. An estimate of total forestland in 1630 
in 19 States in which cerulean warblers occur today and for which there 
was analyzed BBS data (Sauer 2006) was 133,000,000 ha (328,695,000 ac) 
(Smith et al. 2004, p. 33, citing Kellogg 1909). Today, the estimate of 
forest cover in those same States is 73,600,000 ha (181,850,000 ac) 
(Smith et al. 2004, p. 33), a total reduction of approximately 45 
percent. The most dramatic change occurred between the early 1600s and 
1900, when approximately 51 percent of forestland was converted to 
agricultural and other uses (Smith et al. 2004, p. 33). Since 1900, 
approximately 8,500,000 ha (21,000,000 ac) have reverted from primarily 
agricultural uses to forestland. Approximately 52 percent of today's 
hardwood forest within the eastern United States is in mature sawtimber 
(Smith et al. 2004, p. 64); some of this area is northern hardwood 
forest and outside the range of the cerulean warbler.
    The cerulean warbler appears capable of using previously unoccupied 
stands that have matured to develop necessary habitat characteristics. 
Evidence of this capacity comes from New Jersey, New York, and parts of 
New England, where the species has recently expanded its range (Hamel 
1992, pp. 385-400; Robbins et al. 1992, p. 551). Population information 
indicates that this expansion occurred during the later part of the 
1900s, although experts suggest that the expansion does not appear to 
be continuing today (USFWS 2006, Appendix 4, Part II, p. 5). We do not 
know the distribution of cerulean warblers prior to 1966; therefore, we 
do not know whether this expansion is a reoccupation of restored forest 
or true expansion into an area not previously occupied.
    Despite this recent, gradual increase in the total amount of 
forestland, cerulean warbler populations have declined since 1966, 
according to Breeding Bird Survey data. Several hypotheses could 
explain this phenomenon: (1) The amount of forest stands with diverse 
structure continues to decline even though total forestland acres 
increases; (2) local reductions in nesting opportunities in core 
breeding areas are having disproportionate effects at the population 
level; or (3) factors occurring elsewhere in the species annual range 
or not related to nesting opportunities are causing the decline. We 
will discuss each of the first two of these factors in the following 
text, and the third factor in subsequent sections.
    Rangewide data are not available to quantitatively assess the 
amount of or change in habitat with desired characteristics for 
breeding birds. Nevertheless, several pieces of information are 
important for consideration. It takes hundreds of years for hardwood 
forests to naturally achieve complex structure of mature and old-growth 
forests (Hamel 2000, p. 12 citing Widman), which are characteristic of 
stands selected by cerulean warblers for breeding. Much of the 
reversion of agricultural lands to forestland has occurred since the 
early 1900s; therefore, much of the new acreage in forestland remains 
in relatively younger stands that have yet to achieve desired 
structural complexity. We note, however, that stand heterogeneity is 
likely a more important predictor of habitat quality than simply 
looking at stand age, because natural and anthropogenic disturbances 
can create desired stand complexity. Forest management practices, such 
as high-grading, may also affect habitat quality if the largest trees 
in the stand are removed, reducing structural complexity. Fire 
suppression, species-specific tree diseases, and locally or regionally 
high deer densities may also reduce the complexity of forest structure.
    Effects in a relatively small portion of the species' range, 
compared to the species' entire breeding range, could contribute 
disproportionately to the population decline. This has likely happened 
in the past and may happen in the future. Historically, cerulean 
warblers were probably numerous in the bottomland hardwood forests of 
the Mississippi Alluvial Valley. Today, approximately 80 percent of 
forest in this area has been converted to nonforest uses (Brown et al. 
2000, p. 6). Nesting cerulean warblers currently occur only in 
scattered locations within this region. It is important to note that 
most of this loss occurred before the Breeding Bird Survey began in 
1966. Currently, large-scale habitat loss is occurring in the core of 
the species' range, Kentucky and West Virginia, where mountaintop coal 
mining and valley fill operations through 2012 are expected to remove 
567,000 ha (1.4 million ac) of suitable forest habitat (USEPA 2005). 
The total cumulative forest loss from these activities will likely 
eliminate breeding habitat for 10 to 20 percent of the total cerulean 
warbler population currently occurring within that core area. The loss 
of breeding opportunities for birds in this area may have a 
disproportionate effect on the species' total population size.
    The USDA Forest Service has projected forest change to the year 
2050 (Alig and Butler 2004). These projections are based on prior 
trends in forest change, expected market conditions, and no change in 
forest management related policies. Under these conditions, the Forest 
Service expects a slight decline in hardwood forest area. Hardwoods 
will continue to dominate the southeastern United States; however 
hardwood forest area is expected to decline by up to 18 percent by 2050 
(Alig and Butler 2004, pp. 32-33). Maple-beech-birch and oak-hickory 
forests are estimated to decrease by 6 percent and 15 percent, 
respectively (Alig and Butler 2004 p. 18). We note that small portions 
of the hardwood forest area contained within these estimates are 
outside the range of the cerulean warbler; refer to Alig and Butler 
(2004, p. 2) for a map of the forest survey area. We stress that 
changes in acreage or percent of forest landscape in hardwoods are only 
one determinant, and the actual composition and structure of hardwoods 
forests in future landscapes may be equally or more important.
    In summary, a variety of factors has affected the quantity and 
quality of mature and old-growth hardwood forests within the range of 
the cerulean warbler. Overall, habitat loss beginning in the 1600s 
likely precipitated a decline in cerulean warblers; however, the 
conversion of forests stabilized with about 50 percent of forestland 
remaining in the early 1900s. Rangewide cerulean warbler population 
information did not become available until the 1960s; therefore, we do 
not know how the pre-1900s cerulean warbler population size changed as 
a result of this dramatic habitat loss, nor how it may have responded 
to post-1900 forest changes. Beginning in the 1900s, re-growth of 
forests previously converted to agriculture has added potential 
breeding habitat that may be reoccupied when stands achieve the 
characteristics selected for by cerulean warblers, as evidenced today 
in the Northeastern United States.
    2. Reduction in foraging success resulting from decreased prey 
abundance, primarily on the wintering ground in South America:
    Cerulean warblers feed exclusively on insects in North America, and 
on insects and nectar in South America. Availability of these resources 
is critical to an individual bird's survival. Insufficient fat storage 
before spring migration could increase an individual's risk of 
mortality and decrease reproductive success upon return to the breeding 
grounds. Insufficient fat

[[Page 70725]]

storage before fall migration could leave an individual at risk of 
mortality, especially if the migration route is over water where 
foraging opportunities are limited, as is currently hypothesized.
    Winter range--Abundance of food resources in South America has 
likely declined because of the degradation and removal of tropical 
forests. Removal of overstory trees, as forests are cleared and shade-
grown coffee plantations are converted to sun coffee plantations, is 
expected to result in losses of arthropods that are specialized for the 
canopy layers. For example, in Costa Rica, Perfecto (1996, p. 602) 
reported an average of 72 percent of the ants in a tropical forest tree 
canopy to be canopy specialists. However, that we do not know that 
cerulean warblers prey on ants. In a Costa Rican study, Perfecto et al. 
(1996, p. 602) reported similar arthropod diversity in overstory trees 
within shade-grown coffee plantations as within a native forest canopy. 
We do not have figures for arthropod diversity or abundance in the 
Northern Andes, but we expect that conditions may be similar. We do not 
have quantitative information on the differences in nectar resources 
between tropical forest and developed lands.
    Moreno et al. (2006, p. 3) used a climatic and geospatial model to 
predict the potential maximum occurrence of cerulean warbler wintering 
habitat in the narrow elevation zone (500 to 1,500 m (1,650 to 5,000 
ft)) in the Northern Andes and estimated a nearly 60 percent current 
reduction from maximum levels. The remaining habitat is tropical forest 
and shade-coffee plantations. Some field biologists believe that the 
model overestimates habitat availability, and they estimate that less 
than 10 percent remains (Moreno et al. 2006 unpublished report, pp. 3, 
5).
    Most of the loss of tropical forests in the Northern Andes occurred 
within the latter half of the 1900s. Approximately 15 percent of the 
species' modeled potential habitat (Moreno et al. 2006 unpublished 
report, p. 5) is managed under protective status. The effectiveness of 
this protective status for conserving cerulean warblers is uncertain 
because none of the documented cerulean warbler winter occurrences are 
within protected areas (Moreno et al. 2006 unpublished report, p. 5). 
The rate of loss of the remaining tropical forest is likely to be 
decreasing because remnant forests are in steep and inaccessible areas; 
however, removal of portions of the remaining tropical forests 
continues.
    We know that cerulean warblers occupy shade-coffee plantations 
during the non-breeding season, but we do not know whether shade-coffee 
plantations are optimal or sub-optimal habitat because data are not 
available to compare body condition of cerulean warbler on shade-coffee 
plantations with birds occupying tropical forests. In other words, 
presence does not necessarily equate to suitability of these habitats. 
The amount of habitat supplied by shade-coffee plantations is 
diminishing, as some of these plantations are converted to sun-coffee 
plantations that lack the overstory required by wintering cerulean 
warblers (Moreno et al. 2006 unpublished report, p. 2). Cerulean 
warblers are not known, and are highly unlikely, to occur in sun-coffee 
plantations due to the plainly inadequate structure of such vegetation.
    In summary, the population-level effects of habitat loss and 
degradation on forage abundance and foraging success have not been 
quantified. It is reasonable to conclude, however, that a greater than 
60 percent decline of wintering habitat in South America has 
contributed to the approximately 3 percent annual population decline of 
cerulean warblers through reduced forage availability and increased 
competition for remaining food resources.
    Breeding and Post-Fledging Range--Under pre-European settlement 
conditions on the breeding grounds, the hardwood forests of the eastern 
United States were a mosaic of different seral stages (Williams 1989, 
pp. 22-49). Although the forests were predominately mature and old 
growth, patches of younger seral-stage forests occurred within small 
gaps (Lorimer 1989, pp. 565-566). Today, cerulean warblers occur in 
greater relative abundance within landscapes with similar mosaic 
characteristics. Information suggests that cerulean warblers select 
nests sites in stands where canopies are interrupted by small gaps and 
canopy closure is between 65 percent and 85 percent (Hamel 2000, p. 
16). Nests are found in areas with large diameter trees and stands with 
complex canopies, but small patches of seedling-sapling aged trees 
within the mature forest mosaic may provide important habitat for post-
fledging first-year birds.
    Today's mature forest characteristics may not mimic the mosaic 
conditions of original hardwood forest because of alterations in the 
disturbance regimes through fire suppression, dense populations of 
deer, and certain timber harvest methods. The effects of this change in 
forest disturbance regimes on cerulean warblers are not well studied or 
understood. It is possible, however, that the replacement of the 
natural disturbance regime--characterized by frequent, small-scale 
disturbances--with the less-frequent larger-scale disturbances (Lorimer 
1989, pp. 565-566) may not produce understory conditions that favor 
foraging success for post-fledging birds because of the lack of 
interspersed seedling-sapling patches.
    3. Increased predation throughout the species' annual range and 
nest parasitism of cerulean warblers in the breeding grounds, resulting 
from habitat fragmentation:
    Fragmentation of cerulean warbler habitat has occurred throughout 
the species' range. High rates of predation and brood parasitism often 
accompany habitat loss and fragmentation, especially in forested 
landscapes interspersed with agricultural lands and grasslands (Hoover 
and Brittingham 1993, p. 234; Brittingham and Temple 1983, pp. 31-34; 
Faaborg et al. in Martin and Finch 1995, p. 361). Several studies have 
shown low rates of nest success (less than 40 percent) for cerulean 
warblers in areas of fragmented forest within agricultural landscapes 
due to high levels of predation and the presence of nest parasitism 
(Hamel 2000a, p. 4; Roth 2004, p. 43; Varble 2006 p. 3). Direct 
measurements of adult and post fledging mortality due to habitat loss 
and fragmentation during the breeding season on cerulean warblers do 
not exist; however, this phenomenon is well documented with other 
canopy and sub-canopy nesting songbird species. It is reasonable to 
conclude that brood parasitism and predation are exacerbated by habitat 
loss and fragmentation and that this is contributing to the 
approximately 3 percent annual population decline.
    Wintering Range--Effects of habitat loss and fragmentation include 
increased risk of mortality from predation of neotropical migrant 
songbirds in the non-breeding range (Rappole et al. 1989, p. 407; Petit 
et al. in Martin and Finch 1995, pp. 179-180), especially if birds are 
forced to wander outside optimal habitat. Although no studies of 
predation on cerulean warblers in the non-breeding range have been 
conducted, it is reasonable to assume that predation-caused mortality 
of cerulean warblers is similar to that documented for other warbler 
species.
    Approximately 60 to 90 percent of wintering habitat of cerulean 
warblers in South America has been converted to other land uses. This 
loss of habitat has resulted in a highly fragmented landscape. 
Geospatial modeling estimates that fragmentation of this ha