# Endangered and Threatened Wildlife and Plants; 12-Month Finding on a Petition To List Kittlitz's Murrelet as an Endangered or Threatened Species

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/fr%3A2013-24172

## Record

- **Collection:** Federal Register
- **Document type:** Proposed Rule
- **Published:** October 3, 2013
- **Citation:** 78 FR 61764

## Text

DEPARTMENT OF THE INTERIOR
Fish and Wildlife Service
50 CFR Part 17
[Docket No. FWS-R7-ES-2013-0099; 4500030113]
Endangered and Threatened Wildlife and Plants; 12-Month Finding on a Petition To List Kittlitz's Murrelet as an Endangered or Threatened Species

AGENCY:

Fish and Wildlife Service, Interior.

ACTION:

Notice of 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 Kittlitz's murrelet (
Brachyramphus brevirostris
) as an endangered or threatened species and to designate critical habitat under the Endangered Species Act of 1973, as amended (Act). After a review of the best available scientific and commercial information, we find that listing the Kittlitz's murrelet is not warranted at this time. However, we ask the public to submit to us any new information that becomes available concerning threats to the Kittlitz's murrelet or its habitat at any time.

DATES:

The finding announced in this document was made on October 3, 2013.

ADDRESSES:

This finding is available on the Internet at
http://www.regulations.gov
at Docket Number FWS-R7-ES-2013-0099. Supporting documentation we used in preparing this finding is available for public inspection, by appointment, during normal business hours at the U.S. Fish and Wildlife Service, Juneau Fish and Wildlife Field Office, 3000 Vintage Blvd., Suite 201, Juneau, AK 99801. Please submit any new information, materials, comments, or questions concerning this finding to the above street address.

FOR FURTHER INFORMATION CONTACT:

Bill Hanson, Field Supervisor, Juneau Fish and Wildlife Field Office (see
ADDRESSES
); by telephone at 907-780-1160; or by facsimile at 907-586-7099
mailto:.
If you use a telecommunications device for the deaf (TDD), please call the Federal Information Relay Service (FIRS) at 800-877-8339.

SUPPLEMENTARY INFORMATION:

Background

Section 4(b)(3)(B) of the Act (16 U.S.C. 1531
et seq.
), requires that, for any petition to revise the Federal Lists of Endangered and Threatened Wildlife and Plants that contains substantial scientific or commercial information that listing the species may be warranted, we make a finding within 12 months of the date of receipt of the petition. In this finding, we will determine that the petitioned action is: (1) Not warranted, (2) warranted, or (3) warranted, but the immediate proposal of a regulation implementing the petitioned action is precluded by other pending proposals to determine whether species are endangered or threatened, and expeditious progress is being made to add or remove qualified species from the Federal Lists of Endangered and Threatened Wildlife and Plants. Section 4(b)(3)(C) of the Act requires that we treat a petition for which the requested action is found to be warranted but precluded as though resubmitted on the date of such finding, that is, requiring a subsequent finding to be made within 12 months. We must publish these 12-month findings in the
Federal Register
.

Previous Federal Actions

We received a petition dated May 9, 2001, from the Center for Biological Diversity, Coastal Coalition, Eyak Preservation Council, Lynn Canal Conservation, Inc., and Sitka Conservation Society, requesting that the Kittlitz's murrelet be listed as an endangered or threatened species and critical habitat be designated. Included in the petition was supporting information regarding the species' taxonomy and ecology, historical and current distribution, status, and potential causes of decline. We acknowledged receipt of the petition in a letter to the Center for Biological Diversity, dated June 7, 2001. In that letter we stated that, due to funding constraints in Fiscal Year 2001, we would not be able to begin processing the petition at that time, but would request the appropriate funding for Fiscal Year 2002. We also stated that emergency listing of the Kittlitz's murelet was not warranted at that time.

On June 13, 2002, we received a 60-day notice of intent to sue from the Center for Biological Diversity alleging a violation of section 4 of the Act for failure to complete 90-day and 12-month findings on the petition.

On May 4, 2004, we published a candidate notice of review (CNOR) in the
Federal Register
(69 FR 24876) in which the Kittlitz's murrelet was included in the
Summary of New Candidates.
In this document, we indicated that listing of the Kittlitz's murrelet under the Act was warranted but precluded rangewide, and we assigned a listing priority number (LPN) of 5 to this species. The LPN of 5 reflected non-imminent threats of high magnitude for this species. On May 11, 2005 (70 FR 24870) and September 12, 2006 (71 FR 53756), we retained Kittlitz's murrelet in our CNORs with a LPN of 5.

On December 6, 2007, we published an annual CNOR in the
Federal Register
(72 FR 69034) that included a notice of change in LPN for the Kittlitz's murrelet, elevating it from a 5 to a 2 to acknowledge that threats facing this species were of high magnitude and imminent. The CNORs in 2008 (73 FR 75176, December 10, 2008), 2009 (74 FR 57804, November 9, 2009), and 2010 (75 FR 69222, November 10, 2010) continued to assign a LPN of 2 to Kittlitz's murrelet.

On July 12, 2011, the Service reached a multi-district litigation settlement agreement with the Center for Biological Diversity that requires the Service to review and address the needs of over 250 species, including the Kittlitz's murrelet, included in a CNOR published in the
Federal Register
on November 10, 2010 (75 FR 69222). The Kittlitz's murrelet was included in the settlement, requiring the Service to submit a proposed rule or not-warranted finding to the
Federal Register
by September 30, 2013.

On October 26, 2011, the CNOR (76 FR 66370) included a notice of change in listing priority for the Kittlitz's murrelet, downgrading the LPN from a 2 to an 8 because we determined through a reassessment of the threats that their magnitude was moderate (not high) and threats were imminent.

In Fiscal Year 2012, the Service initiated work on the listing evaluation of the Kittlitz's murrelet, as stated in the November 21, 2012 CNOR (77 FR 69994).

This document addresses our requirements under the multi-district litigation settlement agreement.

Species Information

This document constitutes a 12-month finding on the May 9, 2001 petition to list the Kittlitz's murrelet as an endangered or threatened species.

The petitioners requested the Kittlitz's murrelet be listed as an endangered or threatened species and we confirm that this species is a listable entity under the Act. Because we find that listing the Kittlitz's murrelet rangewide is not warranted, as explained below in the Finding section, we conducted further analysis to evaluate any potential distinct population segments (DPS) or significant portion(s) of the range (SPR) within the range of the Kittlitz's murrelet exist that may require listing. However, we did not identify any

populations of the Kittlitz's murrelet that meet the definition of DPS or SPR (see appropriate sections below).

Taxonomy and Species Description

Kittlitz's murrelet (
Brachyramphus brevirostris;
Vigors 1829) is a member of the Alcidae or auk family.
Brachyramphus
murrelets are unusual because unlike the rest of this diverse family of seabirds that nest in colonies, they nest solitarily. There are two additional species within the genus, the marbled murrelet (
B. marmoratus;
Gmelin 1789) and the long-billed murrelet (
B. perdix;
Pallas 1811; Friesen
et al.
1996a, p. 360). The distributions of marbled murrelet and Kittlitz's murrelet overlap in Alaska and the distribution of the long-billed murrelet overlaps with the Kittlitz's murrelet in portions of eastern Russia (Friesen
et al.
1996b, p. 682). All three species generally are similar in appearance, but physical and genetic differences among them are well documented (Pitocchelli
et al.
1995, pp. 239-248; Friesen
et al.
1996a, pp. 363-365; Friesen
et al.
1996b, pp. 681, 685-687; Day
et al.
1999, p. 2). Kittlitz's murrelets are heavier (8.3 ounces [oz] (236 grams [g])) (Kissling, Service, 2007-2012, unpublished data), and have larger heads, longer wings and tails, and smaller bills than do marbled murrelets (7.7 oz [219 g]) (Pitocchelli
et al.
1995, pp. 241-245; Kuletz
et al.
2008, pp. 91-95; Kissling, unpublished data). Long-billed murrelets are distinctly larger than both Kittlitz's and marbled murrelets, have a longer bill than them, and have a white eye ring (Friesen
et al.
1996b, p. 681).

Mitochondrial DNA (mtDNA) sequences and restriction fragment analysis show significant differentiation among the three species (Pitocchelli
et al.
1995, pp. 244-247; Friesen
et al.
1996a, pp. 364-366; Friesen
et al.
1996b, pp. 683-687). Analysis of allozymes further strengthens the evidence that these murrelets are separate species (Friesen
et al.
1996a, pp. 361-365). In addition, nuclear introns and cytochrome
b
gene sequencing showed no evidence of recent hybridization between marbled and Kittlitz's murrelets (Pacheco
et al.
2002, pp. 179-180).

The Kittlitz's murrelet has been considered a single panmictic population (with random mating of individuals within a breeding population) for lack of any evidence to suggest otherwise, but several recent studies suggest that there is strong population genetic structure in this species (MacKinnon 2005, pp. 18, 24-25; Birt
et al.
2011, pp. 47-49; Friesen and Birt 2012, pp. 6-9). Intra-specific analyses of genetic data (allozymes, cytochrome
b
gene, control region of mtDNA, and nuclear DNA) indicate that there are two strongly differentiated genetic groups: one in the western Aleutian Islands and the other in the Gulf of Alaska (Friesen
et al.
1996b, p. 686; MacKinnon 2005, pp. 18, 24-25; Birt
et al.
2011, pp. 47-49; Friesen and Birt 2012, pp. 6-9). Birt
et al.
(2011, pp. 46, 49) concluded that gene flow between these two groups has been very limited for an extended period of time and that the genetic structure probably is due to historical fragmentation of populations; however, this study was based on limited sample sizes within and among populations of Kittlitz's murrelet (53 individuals from three study sites; n=15 from Attu, n=18 from Kachemak Bay, n=20 from Glacier Bay).

Friesen and Birt (2012, pp. 9, 16) expanded the study to include 301 individuals from nine study sites in coastal Alaska ranging from Glacier Bay in the south to Barrow in the north; results supported the previous findings of strong genetic structure in the Kittlitz's murrelet, resulting in an eastern group (Glacier Bay to Kodiak Island) and a western one (Adak, Agattu and Attu islands) that probably diverged from one another a long time ago (547,428 years ago; 95 percent confidence interval [CI]=131,000 to 896,000; confidence intervals are a range of values defined so that there is a specified probability that the value of a parameter lies within it) (Friesen 2013, in litt.). In addition, there are two contact zones where Kittlitz's murrelets have mixed ancestry from both groups; those contact zones are located between Atka and Unalaska islands in the eastern and central Aleutian Islands and in northern Alaska, although the sample size from this latter area was small (n=9) (Friesen and Birt 2012, pp. 10, 16).

Importantly, results from the expanded genetic study suggest that there are low levels of contemporary movement between the two groups and that Kittlitz's murrelets from the two groups can and do interbreed and that offspring are viable and fertile (Friesen and Birt 2012, p. 10). Therefore, birds within the two groupings (eastern and western) do not constitute separate species because genetic connectivity still exists (Friesen and Birt 2012, p. 10). Further, although a comprehensive, comparative study has not occurred yet, there are no documented differences in morphology (e.g., plumage, size) or behavior of Kittlitz's murrelets from the eastern and western genetic groups or across their range (Day
et al.
1999, pp. 2, 20; Day 2013, in litt.). Both groups have sufficient levels of intra-specific genetic variation and do not have evidence of a genetic bottleneck (Friesen and Birt 2012, pp. 17-18; Kissling 2012, in litt.). To date, there have been no genetic analyses comparing Kittlitz's murrelets from Russia with those from North America (preliminary laboratory work has been initiated but was not completed at the time of writing of this finding). We recognize the two genetic groupings (eastern and western), but do not consider these groups to meet the definition of a DPS (see below).

Distribution

The range of the Kittlitz's murrelet encompasses a vast area from the Russian Far East (northern Okhotsk Sea, Bering Sea coast, and coast of the Chukchi Sea in northern Chukotka as far to the northwest as Cape Schmidt) across to the Aleutian Islands and southeastern Alaska, and north to northwestern Alaska (Day
et al.
1999, pp. 3-6; Artukhin
et al.
2011, p. 29). Nests have been recorded throughout nearly the entire at-sea range. Seasonal shifts in distribution are discussed below. There is no reliable information to suggest that the historical range of the Kittlitz's murrelet is substantially different than the current range.

Habitat and Life History

In this section, we describe seasonal shifts in distribution and habitats used, molting cycles, foraging and nesting characteristics, and the demography of the Kittlitz's murrelet.

The Kittlitz's murrelet typically spends its entire annual cycle in marine waters within and adjacent to Alaska and eastern Russia, generally moving offshore (maximum observed 106 miles [mi] (170 kilometers [km]) from shore) during the non-breeding months (August-March or April) and nearshore (within 3.4 mi [5.5 km] from shore) during the breeding season (April-August) with some latitudinal variation. Low numbers of adult Kittlitz's murrelets also have been observed during the breeding season on freshwater lakes (Savage 2013, in litt.; Walsh 2013, in litt.). The seasonal appearance, increase, and disappearance of Kittlitz's murrelets during systematic surveys at sea during the breeding season (Klosiewski and Laing 1994, pp. 55, 83; Kendall and Agler 1998, p. 55; Kuletz
et al.
2003a, pp. 17-20; Robards
et al.
2003, pp. 92, 100, 104; Kissling
et al.
2007, pp. 2167-2168; Kuletz
et al.
2008, pp. 21-22, 53-54) demonstrate that murrelets move inshore near to known breeding areas in south-coastal Alaska beginning in March or April, peak in densities in late

June and early July, and leave these areas rapidly, but asynchronously in late July to mid August. Post-breeding movements of murrelets in late July and August are westward to nearshore waters of Kodiak Island and along the Alaska Peninsula and Bristol Bay, then northward to the Bering and Chukchi seas and even extending, in a few cases, into the Beaufort Sea, where birds may remain until about late October when their pre-basic molt is complete (Day
et al.
2011, pp. 57-59; Madison
et al.
2012, p. 1). At-sea surveys have documented the regular occurrence of Kittlitz's murrelets from August through October in offshore waters between Cape Peirce and north of Nunivak Island, and north of the Bering Strait from Cape Lisburne to the western Beaufort Sea (Kuletz, Service, 2006-2012, unpublished data). In November, as sea ice builds in the Chukchi and Beaufort seas, Kittlitz's murrelets begin to move south into the Bering Sea where they probably winter until late February or early March (Day
et al.
1999, p. 7; Kuletz and Lang 2010, pp. 39-43; Day
et al.
2011, p. 59). However, records of winter sightings in southeastern, south-central, and western Alaska (Klosiewski and Laing 1994, p. 83; Kendall and Agler 1998, pp. 55-56; Day
et al.
1999, pp. 4-5; Day 2006, pp. 208-209; Stenhouse
et al.
2008, p. 61) indicate that some individuals are year-round residents in these areas. Annual movements of Kittlitz's murrelets in eastern Russia, the Aleutian Islands, and northern Alaska remain poorly known, although limited satellite-tag data indicate that Kittlitz's murrelets in the central Aleutians follow the same northward post-breeding migration to the Bering and Chukchi seas as those birds tagged in the Gulf of Alaska do (Madison
et al.
2012, p. 1).

The winter range of the Kittlitz's murrelet is poorly known (Day
et al.
1999, pp. 4-5). Recent information from icebreaker-based at-sea surveys indicates that open water leads (fractures in sea ice caused by wind drift or ocean currents) and polynyas (a large area of open water surrounded by sea ice), primarily south of St. Lawrence Island, between Nunivak and St. Matthew islands, and east of the Pribilof Islands, may be important wintering areas (Kuletz and Lang 2010, pp. 40-43; Kuletz, unpublished data). Most Kittlitz's murrelets encountered during early spring surveys in the sea ice were in pairs (Kuletz and Lang 2010, p. 40). The exact winter distribution of Kittlitz's murrelets in the Bering Sea probably shifts with respect to dynamic changes in open leads and polynyas (Kuletz, unpublished data), which tend to form consistently near the large Bering Sea islands and some coastal areas (Niebauer
et al.
1999, p. 34). The winter range of the species in eastern Russia is largely unknown, but observations have been reported from the Kamchatka Peninsula and the Kuril Islands in the Russian Far East south to northern Japan (Flint
et al.
1984, pp. 156-157; Brazil 1991, p. 164; but see Carter
et al.
2011, p. 8). A few birds also have been observed during late winter in the Sireniki polynya of southern Chukotka and the western Bering Sea in Russia (Konyukhov
et al.
1998, p. 325; Shuntov 2000, pp. 97-98).

During the summer breeding season, Kittlitz's murrelets usually, but not exclusively, are associated with glacially influenced waters, especially those with floating ice, in south-coastal Alaska, where large numbers aggregate (Isleib and Kessel 1973, p. 100; Kendall and Agler 1998, p. 58; Day
et al.
2000, p. 109; Arimitsu
et al.
2011, p. 18; Hoekman
et al.
2011, p. 40; Kissling
et al.
2011, p. 7; Kuletz
et al.
2011a, pp. 102-103; Kuletz
et al.
2011b, pp. 90-92; Piatt
et al.
2011, p. 70; Arimitsu
et al.
2012, p. 18). The exact reasons for this association are unclear, but hypothetical explanations exist. This pattern of at-sea distribution simply may reflect an adaptation for nesting on unvegetated scree slopes or nunataks (isolated peaks of rock projecting above the surface of inland snow or ice) often associated with tidewater glaciers, which are selected because these areas are thought to be predator-free (Piatt
et al.
1999, p. 12; Kissling
et al.
2012, p. 1; Lawonn 2012, pp. 21, 94-95). Their association with tidewater glaciers also may reflect foraging preference and efficiency in glacial-affected water (Day
et al.
2003, pp. 681, 686; Kuletz
et al.
2003b, p. 138; Allyn
et al.
2012, pp. 244-245; Arimitsu
et al.
2012, pp. 14, 18). In addition, strong nest area and site fidelity may cause these birds to return to the same area (Piatt
et al.
1999, p. 11; Kaler
et al.
2010, p. 18; Lawonn 2012, pp. 82, 88; Kenney and Kaler 2013, p. 73; Kissling, unpublished data), but it is unknown if the same birds are using a particular area annually or if site characteristics make the area suitable to breeding pairs.

The Kittlitz's murrelet has two distinct plumages in its annual cycle and therefore undergoes two molts per year: a full, pre-basic molt in fall (September-October) and a partial, pre-alternate molt in spring (April-May) (Day
et al.
1999, pp. 18-19). During the pre-basic molt, individuals transition from their mottled, cryptic plumage of the breeding season to the sharply contrasting black and white plumage of the non-breeding season. The pre-basic molt replaces of the wing, tail, and body feathers, whereas the pre-alternate molt replaces only the body feathers. Although Sealy (1977, p. 467) reported that in the pre-basic molt wing feathers grow synchronously rendering a flightless period (2-4 weeks) for the bird, Pyle (2009, p. 222) found that Kittlitz's murrelets undergo a non-synchronous molt, either sequentially or in blocks, perhaps to avoid an extended flightless period, and probably prolonging the pre-basic molt period.

Foraging

Because little research on the Kittlitz's murrelet has occurred during the winter, information about foraging and other life-history characteristics are based primarily on observations made during the spring, summer, and fall. Kittlitz's murrelets tend to forage as single birds or in small groups, but seldom in mixed-species feeding flocks (Day and Nigro 2000, pp. 8-10, 12). Most foraging occurs during the day (Day
et al.
1999, p. 9; Madison
et al.
2010, p. 1), especially in the morning (Day and Nigro 2000, p. 5). They pursue and capture prey underwater by using wing-propelled “flight” and consume prey either at the surface or underwater (Day
et al.
1999, p. 9; Day and Nigro 2000, p. 9).

The Kittlitz's murrelet appears to be a flexible forager with a diet that varies considerably among seasons but is fairly specialized within a season (Hatch 2011, pp. 25-26, 35; Allyn 2012, p. 102). Although Kittlitz's murrelets are considered to be piscivorous, they also eat zooplankton throughout the entire annual cycle (Day
et al.
1999, p. 9), more so than for the closely related marbled murrelet (Hobson
et al.
1994, p. 795; but see Day
et al.
1999, p. 10). In the pre-breeding season, Kittlitz's murrelets feed on low-trophic-level prey such as macrozooplankton and larval fishes and gradually transition to consuming larger proportions of higher-trophic-level prey (planktivorous fishes) as the breeding season commences (Hatch 2011, pp. 24-25; Allyn 2012, p. 102). During the breeding season, Kittlitz's murrelets feed on a combination of macrozooplankton (36-44 percent of their diet) and schooling fishes such as Pacific capelin (
Mallotus villosus
), Pacific sand lance (
Ammodytes hexapterus
), juvenile Pacific herring (
Clupea pallasi
), and juvenile walleye pollock (
Theragra chalcogramma
) (Sanger 1983, p. 692; Hobson
et al.
1994, p. 795; Day
et al.
1999, p. 9; Day and Nigro 2000, pp. 11-13; Kuletz
et al.
2003a, pp. 23, 28; Agness 2006, p. 119; Kuletz
et al.
2008, p. 26; Hatch 2011, p. 47; Kaler
et al.

2011, p. 15; Allyn 2012, p. 102; Lawonn 2012, pp. 27-28). By the post-breeding period they feed almost exclusively on these high-lipid fish, consuming only small proportions (4-9 percent) of zooplankton (Hatch 2011, p. 47; Allyn 2012, pp. 100-101). In the northern Bering and Chukchi seas, a variety of small arctic fishes and large zooplankton are abundant (Eisner
et al.
2013, pp. 97-102) and presumably are consumed by Kittlitz's murrelets in the fall and winter. Based on a comparison of stable isotopes (carbon and nitrogen) from recently captured murrelets and museum specimens, these seasonal foraging patterns have been consistent over the past century (1911-2009) (Hatch 2011, p. 27).

During nesting, Kittlitz's murrelets carry a single whole fish at a time to their chick. Adult fish-holding murrelets often stage on the water before returning to their nest to deliver the fish to the chick; therefore, chick diet has been inferred by identifying these fishes held in the bill of adults on the water and by directly monitoring food deliveries to chicks at nest sites. The proportion of fish held in bill by adults on the water that is identified to species is low (21-23 percent) (Agness 2006, p. 116; Kuletz
et al.
2008, p. 26) because of the difficulty for the observer to do so at a distance and under at-sea conditions. This method is useful, however, in areas where it is difficult to monitor nests directly such as in glacial-dominated landscapes, where Kittlitz's murrelets have been observed on the water holding primarily sand lance and capelin, and to a lesser extent Pacific herring and Pacific salmon (
Oncorhynchus
spp.) (Agness 2006, p. 124; Kuletz
et al.
2008, p. 26). In contrast to the low identification rate of fish held by murrelets on the water, most fish (70-85 percent) delivered to chicks at monitored nests have been identified to species (Naslund
et al.
1994, p. 46; Lawonn 2012, p. 27-28; Kaler 2012, in litt.; Kissling, unpublished data). Pacific sand lance is the fish species delivered most commonly to chicks (57 percent of identified deliveries) and occurs in chick diet in all areas where nests have been monitored (n=33 nests; western Aleutians and Kodiak islands and Kachemak and Icy bays) (Naslund
et al.
1994, p. 46; Lawonn 2011, pp. 27-28; Kaler 2012, in litt.; Kissling, unpublished data). Although significant geographic variation exists (see
Nesting,
below), the remainder of chick diet is composed of hexagrammids (23 percent; kelp greenling [
Hexagrammos decagrammus
] and Atka mackerel [
Pleurogrammus monopterygius
]), capelin (10 percent), gadids (5 percent; Pacific cod [
Gadus macrocephalus
]) and rockfish (
Sebastes
spp.), smelt (2 percent; Osmeridae family) and Pacific herring (1 percent) (Naslund
et al.
1994, p. 46; Lawonn 2011, pp. 27-28; Kaler 2012, in litt.; Kissling, unpublished data). In both methods used to determine chick diet, it is not known if there is bias associated with fish identification due to size of the prey item, but this is certainly possible.

Small schooling fishes that are oily, such as sand lance and capelin, are thought to be favored for chick meals because of their high lipid, and therefore energy, content (van Pelt
et al.
1997, p. 1395; Anthony
et al.
2000, p. 75; Litzow
et al.
2004, p. 1150). Capelin, in particular, is hypothesized to be an important prey species for Kittlitz's murrelets in glacially-affected waters because this fish species occurs in cold, turbid marine waters close to tidewater glaciers (Arimitsu
et al.
2008, p. 137). Chicks eating oily fishes receive more calories and grow faster (Ostrand
et al.
2004, p. 69), resulting in fewer foraging trips for parents, when high-energy fishes are fed to chicks than when lower-energy fishes such as walleye pollock or rockfishes, are fed to chicks (Hatch 2011, pp. 74-77, 103-104). Therefore, a change in the availability of high-energy forage fishes during the breeding season could affect the reproductive success of Kittlitz's murrelets (van Pelt
et al.
1997, p. 1393; Anderson and Piatt 1999, p. 117; Becker
et al.
2007, pp. 276-278; Österblom
et al.
2008, pp. 967-974).

Several studies have described marine habitat use of Kittlitz's murrelets in the breeding season by associating murrelet distribution with marine biotic and abiotic factors in areas where glaciers exist (Day and Nigro 2000, pp. 8-9; Day
et al.
2003, pp. 685-694; Kissling
et al.
2007, p. 2168; Kuletz
et al.
2008, p. 24-27; Allyn
et al.
2012, pp. 240-242; Arimitsu
et al.
2012; pp. 12-14; Renner
et al.
2012, pp. 2035-2039). Generally, Kittlitz's murrelets prefer to forage in shallow (less than 196 feet [ft] (60 meters [m])), glacially affected waters (Kuletz
et al.
2008, p. 37) often with some floating ice (Day and Nigro 2000, pp. 6, 8; Day
et al.
2003, pp. 686, 694; Kuletz
et al.
2003b, pp. 136, 139), but it is not known whether ice occurrence is biologically meaningful to murrelets. Arimitsu
et al.
(2012, p. 18) postulated that the presence of ice may instead serve as a proxy to other factor(s), such as outflow of sediment-laden freshwater from glacial streams and a downstream increase in the availability of certain near-surface prey (e.g., euphausiids). Kuletz
et al.
(2003b, p. 139) hypothesized that the undersides of icebergs and pack ice may increase prey abundance and availability to murrelets, perhaps due to the presence of sea ice algae and its role in primary production (Grebmeier
et al.
2006, p. 339). Other studies have positively associated Kittlitz's murrelets with highly turbid waters (Day
et al.
2003, p. 685; Renner
et al.
2012, pp. 2038-2039), often with a clear, cold freshwater lens at the surface less than 32 ft (10 m) in depth (Kuletz
et al.
2008, p. 37; Allyn
et al.
2012, p. 233); in fact, Day
et al.
(2003, p. 695) suggest that the eyes of Kittlitz's murrelets are large to increase their ability to forage in highly turbid water. This species prefers marine waters with sea surface temperatures of 37-48 degrees Fahrenheit (F) (3-6 degrees Celsius) (Day
et al.
2003, p. 685; Day
et al.
2011, p. 59; Allyn
et al.
2012, p. 242). Kittlitz's murrelets are often associated with areas of localized upwelling that are generally created by the interaction of landscape features, such as submerged marine sills, shoreline, hanging and tidewater glaciers, and strong tidal currents (Day and Nigro 2000, p. 5; Kuletz
et al.
2003b, p. 139; Kissling
et al.
2007, p. 2171; Allyn
et al.
2012, pp. 244-245; Arimitsu
et al.
2012, p. 10), but not tidal height (Allyn 2012, p. 101). It is not known to what extent the distribution of Kittlitz's murrelet depends on these marine habitat conditions for foraging efficiency or prey availability in a given year. However, it is logical to assume that daily, weekly, monthly and annual variability in Kittlitz's murrelet population density at a location may be due, at least in part, to corresponding variability in prey abundance.

Kittlitz's murrelets probably switch among prey types between seasons or years depending on availability, as do marbled murrelets (Ostrand
et al.
2004, p. 73; Becker
et al.
2007, p. 274). High-lipid forage fishes are expected to represent higher-quality prey for seabirds than are zooplankton because the fishes' larger size should result in more energy gained per unit of effort spent foraging (Norris
et al.
2007, p. 876), although macrozooplankton are not necessarily of lower caloric value than fishes (Vermeer and Cullen 1982, p. 35; Davis
et al.
1998, p. 151; Hedd
et al.
2002, pp. 229-230). Janssen
et al.
(2009, p. 36) reported that in some years female marbled murrelets producing eggs early in the breeding period had a higher proportion of low-trophic-level prey in the pre-breeding diet than did murrelets not producing eggs, signifying that low-trophic-level prey may not

necessarily equate to low-quality prey. Thus, the substantial amount of marine invertebrates in the Kittlitz's murrelet diet coupled with their prey-switching abilities, may buffer self-feeding adults from annual and seasonal variation in the availability or quality of high-energy forage fishes (Anderson and Piatt 1999, p. 117; Robards
et al.
2003, p. 2; Litzow
et al.
2004, p. 1149; Arimitsu 2009, pp. 33-36, 45).

Nesting

The Kittlitz's murrelet is a dispersed-nesting seabird (i.e., does not nest in colonies like most marine foragers) that often nests in remote, rugged areas and therefore little information on their nesting ecology existed until recently. Until 1999, only 19 confirmed Kittlitz's murrelet nests had been described, 17 in Alaska and 2 in Russia (Day
et al.
1999, pp. 25-26). In 2005, a nest was opportunistically discovered on Agattu Island, at the western end of the Aleutian Islands (Kaler 2006, p. 3). Since that time, a number of different studies have been initiated, owing to increasing interest in their conservation status, that have greatly added to our knowledge about the nesting and breeding behavior of this species. On Agattu Island, an additional 86 active nests have been found and monitored (Kaler, Service, 2008-2011, unpublished data), 9 nests have been found on Adak Island (Kenney 2012, in litt.; Kenney and Kaler 2013, p. 74), 75 have been found on Kodiak Island (Corcoran and Mackey, 2012, p. 1; Lawonn 2012, p. 10), and 35 have been found in the glaciated landscape around Icy Bay (Kissling, unpublished data). To date, 234 Kittlitz's murrelet nests have been found in Alaska (n=230; 98 percent) and Russia (n=4; 2 percent) (Felis, U.S. Geological Survey, 2013, unpublished data).

Based on these recent efforts, some generalities can be made about nesting habitat and nest site selection of the Kittlitz's murrelet. Their nesting habitat is characterized by sparsely vegetated or unvegetated scree-fields, talus slopes, barren ground, and cliff and rock ledges in the coastal uplands and mountains, often in the vicinity of glaciers or in historically-glaciated areas (Day
et al.
1983, pp. 267-269; Day 1995, pp. 271-273; Konyukhov
et al.
1998, p. 322; Piatt
et al.
1999, p. 8; Kaler
et al.
2009, p. 366; Lawonn 2012, pp. 83-87; Kissling, unpublished data). Rangewide, barren areas, which are characterized by bare rock, gravel, sand, silt or clay with little or no “green” vegetation present appear to be the preferred nesting habitat the Kittlitz's murrelet owing to disproportionate use relative to availability (Kaler
et al.
2009, p. 366; Lawonn 2012, pp. 90, 101-102; Felis, unpublished data; Kissling, unpublished data). In parts of this species' range, such as Kodiak Island, where mammalian predators exist, the Kittlitz's murrelet appears to avoid nesting near vegetated edges (Lawonn 2012, pp. 90, 101). Dwarf shrub and herbaceous habitats occasionally are used by nesting Kittlitz's murrelets, especially in the Aleutian Islands where nests are positively associated with orange crustose lichens (
Xanthoria
spp.) (Kaler
et al.
2009, p. 366; Kenney and Kaler 2013, pp. 73-74), and where this habitat type is abundant. Nesting habitat of the Kittlitz's murrelet is located adjacent to or associated with glaciers and persistent snow only in south-coastal Alaska where these land cover classes currently exist. Generally, the amount of vegetative cover within a 25-m radius of nest sites is least in glaciated areas of south-coastal Alaska (3 percent) (Kissling, unpublished data), moderate on Kodiak Island (9 percent) (Lawonn, p. 102) and northern Alaska (14 percent) (Felis, unpublished data) and greatest in the Aleutian Islands (51 percent) (Kaler
et al.
2009, p. 366). Despite variation in percent of vegetative cover near nests among these study sites, Kittlitz's murrelets consistently nest in the least vegetated areas available on the landscape within a particular area (Lawonn 2012, p. 90; Kaler, unpublished data; Kissling, unpublished data), presumably to maximize the safety of the nest from predators.

Although the amount of vegetative cover appears to drive nest site selection for the Kittlitz's murrelet both within areas and across their range, other characteristics may also be important. Many of these factors, such as elevation, slope, distance to ocean, aspect, substrate, and local climate, however, often are correlated with low vegetative cover. For example, unvegetated or sparsely vegetated areas tend to occur at higher elevations and on steeper, windward-facing slopes. Moreover, the variation in these attributes across the species' range makes it difficult to draw generalizations about their importance. For example, nests have been found from 0.1 to 45.7 mi (0.2 to 73.5 km) from the ocean, on slopes 0-66 degrees, and at elevations between 419 and 7,378 ft (128 and 2,249 m) above sea level. In general, nests located on the steepest slopes and at the highest elevations occur in south-coastal Alaska, whereas those farthest from the ocean are located in northern Alaska (Felis, unpublished data), but this may reflect overall differences in habitat available. Nest orientation is similarly uninformative at the rangewide scale; based on 196 nests with documented aspect, 50 (26 percent) faced north, 56 (29 percent) faced east, 40 (20 percent) faced south and 50 (26 percent) faced west (Day
et al.
1999, 25-26; Lawonn 2012, p. 84; Kaler, unpublished data; Kissling, unpublished data; summarized by Felis, unpublished data), suggesting that nest site aspect is not relevant (Kaler
et al.
2009, p. 366) or is locally driven (e.g., on Kodiak Island) (Lawonn 2012, pp. 83-84). The importance of small- and medium-sized rocks (roughly 2.0- 11.8 inches [in]) (5-30 centimers [cm]) at and near nests has been reported at several study sites (Day
et al.
1983, p. 267; Kaler
et al.
2009, p. 366; Lawonn 2012, p. 89; Kissling, unpublished data).

Similar to that of the marbled murrelet, much of the behavior and life history of the Kittlitz's murrelet appears to have evolved around predator avoidance, particularly during nesting (Nelson and Hamer 1995, p. 66). While most alcids avoid predators by nesting in inaccessible areas (burrows, crevices) or on open rock ledges and protect their young by nesting in large colonies or by guarding them, the Kittlitz's murrelet places its nest in habitats expected to support low numbers of predators, disperses nests across the landscape, and relies on cryptic coloration and behavior to avoid predator detection. On the mainland in south-coastal Alaska, nunataks appear to be favorable habitats presumably because of their isolation from terrestrial predators (Kissling, unpublished data). On Kodiak Island, the median within-year nearest neighbor distance was found to be 1,128 ft (range=42-5,085 ft) (344 m; range=13-1,550 m) (Lawonn 2012, p. 83). In addition to site selection, murrelets have a variety of morphological and behavioral characteristics to minimize detection by potential predators (summarized by Nelson and Hamer 1995, p. 66).

A single egg is laid in a nest scrape composed of sand- and pebble-sized rocks (more typical in northern Gulf of Alaska) or plant matter (moss and lichens; common in western Aleutian Islands) at the base of a large rock or on a cliff ledge (Day
et al.
1983, p. 267; Piatt
et al.
1994, p. 55; Piatt
et al.
1999, p. 11; Day 1995, pp. 271-273; Kaler
et al.
2009, p. 366; Lawonn 2012, pp. 81-82; Kaler 2012, in litt.; Kenney and Kaler 2013, p. 73; Kissling, unpublished data). The egg is colored pale-green, olive-green, or blue-green with brown mottling, ranging from speckling to streaking (Day
et al.
1983, pp. 265-266; Piatt
et al.
1994, p. 55; Kaler
et al.
2009,

p. 367). Across their range and within areas, egg laying is highly asynchronous, with records ranging from 6 May through 17 July (Day 1996, p. 435; Kaler
et al.
2009, pp. 366-367; Corcoran and Mackey 2012, p. 10; Lawonn 2012, p. 21; Kissling, unpublished data). There is some evidence that Kittlitz's murrelets attempt to renest when a nest fails (Kaler and Kenney 2008, p. 16; Kenney and Kaler 2013, p. 73; Kissling, unpublished data).

The duration of incubation is approximately 30 days (Day
et al.
1999, p. 14; Kaler
et al.
2009, p. 365). Both parents incubate the egg, and loss of a parent can mean failure of the nest (Kissling, unpublished data). Mean hatching dates range from 6 July in Icy Bay (Kissling, unpublished data), to 8 July on Kodiak Island (Corcoran and Mackey 2012, pp. 10-11; Lawonn 2012, pp. 21, 47), and to 17 July on Agattu Island (Kaler, unpublished data); these are consistent with the known or expected hatching dates by geographic region presented by Day
et al.
(1996, p. 435), which range from 14 June in southeastern Alaska to 28 July in the Chukchi Sea. Like the marbled murrelet, Kittlitz's murrelet chicks are semiprecocial and are brooded for approximately 48 hours (Nelson and Hamer 1995, p. 66; Lawonn 2012, pp. 23-24). This short period of brooding requires that thermoregulatory capability be developed quickly after hatching so that the chick can remain unattended and have minimal parental care other than food deliveries.

The chick is fed fish for 21-40 days post-hatch at a rate of 1-12 times per day with considerable variation among individual nests, study areas, and years (Day
et al.
1999, p. 15; Kaler
et al.
2011, p.15; Lawonn 2012, p. 51; Kissling, unpublished data). Both adults feed the chick throughout the day and night (Day
et al.
1999, p. 15; Kaler
et al.
2011, p. 16; Kissling, unpublished data), but most meal deliveries occur in the early morning within a 4-hour period around sunrise (Lawonn 2012, p. 26). Similar to those of the marbled murrelet, Kittlitz's murrelet chicks maintain their camouflaging down until just prior to fledging (Nelson and Hamer 1995, p. 60; Kaler
et al.
2009, p. 367). When they fledge, chicks are 40-60 percent of adult body mass, but their wing length is nearly adult-sized (Day
et al.
1983, p. 272; Kaler
et al.
2009, pp. 368-369; Lawonn 2012, p. 60). Their initial flight from the nest to the ocean can be short from island nests (Kaler
et al.
2009, p. 371; Lawonn 2012, p. 101), or much longer from mainland nests that have been recorded as far as 46 mi (74 km) from the ocean (Day
et al.
1983, p. 272). Russian scientists have speculated that newly-fledged Kittlitz's murrelets stage on upland glacial lakes before departing for the ocean, but this hypothesis has not been substantiated (Kuletz
et al.
2008, p. 13), although low numbers of adult Kittlitz's murrelets have been observed on freshwater lakes during the breeding season (Savage 2013, in litt.; Walsh 2013, in litt.). There also is the possibility that fledglings fly downslope to the nearest river from an inland site and use the river as transportation or orientation to the ocean, but this behavior has not been documented (Day
et al.
1983, p. 272).

Demography

Although demographic data are sparse, Kittlitz's murrelets exhibit life-history characteristics that are similar to other alcids, such as fairly long lifespan (assumed to be approximately 15 years), delayed reproductive maturity (assumed to be approximately 3 years of age), intermittent breeding (i.e., they do not appear to breed annually), and low rates of reproduction (Bessinger 1995, p. 385; De Santo and Nelson 1995, pp. 36-37; Begon
et al.
1996, pp. 494-496; Day
et al.
1999, p. 16; Gaston 2004, pp. 164-167). This life-history strategy depends on the survival of at least a few offspring and recruitment of those offspring into the adult breeding population to maintain population stability. Generally, for a species with this life history strategy, changes in mortality rates of reproductively capable adults have greater population-level effects compared to those of juvenile or sub-adult birds and to changes in reproductive rates; in contrast, for a species that is shorter lived, produces more offspring, and matures at an earlier age, changes in reproductive rates tend to drive population-level effects (Ricklefs 1977, p. 467-468; Roff 1992, p. 45; Beissinger 1995, p. 390).

Reproductive Performance.
Assessing reproductive effort and performance of the Kittlitz's murrelet is particularly challenging because of their non-colonial and purposefully cryptic nesting behavior. Low reproductive success has been both suggested (Day and Nigro 2004, pp. 91-94) and documented in Kittlitz's murrelets (Kaler
et al.
2009, p. 369; Lawonn 2012, pp. 29-30; Kaler, unpublished data; Kissling, unpublished data). Because nesting behavior and nesting success have been monitored for only a few years (since 2006) and only in a few locations, it is unclear whether this low rate of nesting success is typical for the Kittlitz's murrelet, a species in which a breeding pair needs to produce offspring only infrequently, or whether one or more environmental parameters have changed, causing decreased breeding effort or increased egg and chick mortality.

In total, 206 active nests of the Kittlitz's murrelet have been monitored, nearly all of which were discovered as part of studies initiated since 2006 on Agattu and Kodiak islands and Icy Bay (south-coastal Alaska) (Naslund
et al.
1994, p. 46; Kaler
et al.
2009, p. 363; Lawonn 2012, p. 10; Corcoran and Mackey 2012, p. 1; Kenney 2012, in litt.; Kaler, unpublished data; Kissling, unpublished data). The majority of these nests (74 percent) failed; only 23 percent successfully fledging a chick; the nest fate was not able to be determined at 3 percent of the nests. Overall, most of the nest failures were attributed to depredation of the egg or chick (31 percent) and death of the chick (starvation, exposure or disease; 29 percent), followed by unknown cause (21 percent), abandonment (14 percent), accident (3 percent), and parent mortality (2 percent). When analyzed collectively, estimates of daily nest survival (± standard error [SE]; standard error is a measure of variability in the data) at the three locations where nests where regularly monitored were slightly higher in Icy Bay (0.979±0.005) than at Kodiak and the Aleutian islands (0.968±0.003) (see
Factor A
discussion for more details on this analysis). Across the 55-day nesting period, these daily nest-survival rates extrapolate to estimates of nesting success of 0.307 and 0.166, respectively. Nest observations from the three locations where nests were regularly monitored are summarized below, as well as observations of juveniles at sea.

Aleutian Islands
—Since 2005, 96 active Kittlitz's murrelet nests have been found in the Aleutian Islands (Agattu=87 and Adak=9) (Kaler
et al.
2009, p. 366; Kenney 2012, in litt.; Kenney and Kaler 2013, p. 74; Kaler, unpublished data). Nests were found using searches conducted on foot owing to the low, scrubby vegetation and rolling hills (Kenney and Kaler 2013, pp. 73-74). From 95 nests of known fate, 18 chicks successfully fledged (19 percent apparent nesting success; range among years=6-44 percent) (Kenney 2012, in litt.; Kaler, unpublished data), and the fate of one chick was unknown because researchers left the island before its fate was determined (Kaler
et al.
2009, p. 369). Of the 77 failed nests, the apparent cause of nest failure was chick death due to starvation and exposure (40 percent), depredation of egg or chick (25 percent), unknown cause (21 percent), failure of eggs to

hatch followed by abandonment (12 percent), or accident (2 percent). Kaler
et al.
(2011, p. 17) could not definitively assign the causes of chick mortality to either exposure or starvation because multiple factors including diet, weather, and provisioning rates by adults were likely contributors. Fledglings in the Aleutian Islands were approximately 50 percent of the adult body mass (Kaler
et al.
2009, pp. 368, 370-371). This percentage is lower than that calculated for marbled murrelets (58-70 percent) (Kuletz and Marks 1997, p. 423; Nelson and Hamer 1995, p. 60; Kissling, unpublished data) and for Kittlitz's murrelets that fledged from Kodiak Island (58 percent) (Lawonn 2012, p. 60) and Icy Bay (63 percent) (Kissling, unpublished data), but is greater than a Kittlitz's murrelet fledgling found on the Kenai Peninsula (40 percent) (Day
et al.
1983, p. 272). The low fledging weight in the Aleutian Islands was most likely due to the poor quality (i.e., low lipid content) of prey delivered to chicks, which included mostly hexagrammids (40 percent of deliveries at 10 nests monitored), sand lance (36 percent), and gadids and rockfish (24 percent), and was reflected in the high prey delivery rates at nests in the Aleutians (9.8 fish per day) (Kaler, unpublished data), which was nearly double the rates observed in the northern Gulf of Alaska (Lawonn 2012, pp. 27, 55; Kissling, unpublished data).

South-central Alaska
—In 1994, one active Kittlitz's murrelet nest was opportunistically found and monitored using a remote video camera on Red Mountain near Kachemak Bay (Naslund
et al.
1994, p. 46; Piatt
et al.
1994, p. 55). The chick fledged and the nest was deemed to be successful (Naslund
et al.
1994, p. 46).

In 2006, an active nest that contained a live Kittlitz's murrelet nestling was found opportunistically on Kodiak Island, although the fate of this nest was not confirmed (Stenhouse
et al.
2008, p. 59). Since then, 74 additional nests have been found by systematically searching areas of apparently suitable habitat on foot in a pre-defined study area (Corcoran and Mackey 2012, p. 1; Lawonn 2012, p. 21). Of these 74 nests, 16 chicks fledged from 71 nests (23 percent apparent nesting success) and nest fate of 3 nests was unknown. The overall annual nest survival rate (number of chicks fledged per nesting pair) was 0.0933 (95 percent CI = 0.0067 to 0.2991) between 2008 and 2011, almost certainly below 30 percent nesting success (Lawonn 2012, p. 30). Nest failures were most commonly caused by depredation (54 percent), followed by chick death (25 percent) and abandonment (20 percent); one nest failed for unknown reasons (Corcoran and Mackey 2012, p. 3; Lawonn 2012, p. 59). The red fox (
Vulpes vulpes
) was the only identified nest predator (13 of 15 predation events recorded; two unidentified predators) (Corcoran and Mackey 2012, p. 3; Lawonn 2012, pp. 30-31). In 2011 and 2012, nine dead chicks found in nest scrapes of monitored nests were necropsied, and all were in fair to good body condition, suggesting that nutritional health was not responsible for their death (Shearn-Bochsler
et al.
2013, p. 1). However, at least six of these chicks had high levels of saxitoxin, a neurotoxin produced by certain species of dinoflagellates, in their gut and/or liver, which is believed to have caused the death of these chicks immediately after consuming sand lance (Shearn-Bochsler
et al.
2013, p. 1). Chick meal delivery rates (±1 standard deviation [SD]; standard deviation is a measure of variability in the data) averaged 4.6 (±0.8) fish per day or 117 (±37) fish from hatching to fledging of the chicks with sand lance being the most common prey delivered (92 percent of deliveries), followed by capelin (8 percent) and a few herring and salmonids (
Oncorhynchus
spp.) (Lawonn 2012, pp. 27-28, 55). On Kodiak Island, the mean number of days to fledging (±SD) was 24.8 (±2.3) days (Lawonn 2012, p. 55), or lower than that for nests monitored at Agattu Island (30.6±5.6 days) (Kaler, unpublished data), despite comparable apparent nesting success at these study sites where similar methods were used to locate and monitor Kittlitz's murrelet nests.

Southeastern Alaska
—In contrast to Kodiak, Adak, and Agattu islands, the terrain in southeastern Alaska is characterized by steep mountains, icefields, and glacial fjords usually with thick vegetation along the near shore areas precluding nest searching efforts by foot. Thus, from 2007 to 2012, 35 Kittlitz's murrelet nests have been located in Icy Bay by tracking 24-44 radio-marked birds throughout each of the six breeding seasons (Kissling, unpublished data). Thus, this is the only study site where some reproductive measures, such as breeding propensity and adult body condition prior to breeding, are available and where nest locations are seemingly unbiased because all habitats within the study area were available to the marked birds for nesting (as opposed to searching a specified area that consists of presumably suitable nesting habitat). The mean proportion of radio-marked murrelets that attempted to nest annually was 0.18 (range=0.03-0.43 across all years), but because weather and logistics precluded daily aerial tracking, it is possible that early failed breeders were not detected and that this estimate of breeding propensity is biased low. Therefore, Kissling (unpublished data) used a combination approach to estimate breeding propensity that includes quantifying levels of vitellogenin (an egg-yolk precursor protein expressed only in females), brood patch development (necessary for incubation in both sexes), and radio-telemetry (following Peery and Henry 2010, p. 2417). Using the combination method, the proportion of murrelets attempting to breed was 0.87 (range=0.75-1.00), which is probably biased high because brood patches can be an unreliable indicator of reproductive status (McFarlane Tranquilla
et al.
2003, p. 112). It is difficult to reconcile the range in estimates of breeding propensity (0.18-0.87; mean=0.52; breeding propensity is defined as the probability that an after-second-year murrelet will breed in a given year), and it is impossible to determine the accuracy of either method because in glacial-dominated landscapes such as Icy Bay, alternative field methods to locate nests do not currently exist. Many adult Kittlitz's murrelets arrive in Icy Bay paired with a mate and in apparently good body condition, suggesting perhaps that certain environmental cues may be required for breeding to proceed. Another possible explanation for the variable breeding-propensity rate is that there is a capture, handling, or radio-transmitter effect on individual Kittlitz's murrelets; however, several lines of evidence, including few juveniles observed at sea and good reproductive performance of radio-marked marbled murrelets (see below), suggest that this possible issue is not significant.

Because most (86 percent) nests in Icy Bay were not accessible due to the dangerous terrain, nest fate often was inferred (following Bradley
et al.
2004, pp. 321-322), but nests occasionally (n=5) could be monitored with video or still cameras. Apparent nesting success across all years combined was 40 percent (14 of 35 nests). Causes of failure were largely unknown (71 percent of failed nests) because most of the nests were inaccessible, but of those where cause of failure could be determined or inferred, three failed due to parent mortality (predation) during incubation, two failed due to unstable terrain (i.e., a rockfall), and one egg was abandoned. Despite the small sample sizes, successful nests (n=14) were

located closer to the ocean (median distance=5.6 mi [9.0 km]) than failed nests (n=21; median distance=15.0 mi [24.1 km]); the elevation of nests did not affect nest fate (4,226 ft [1,288 m] for successful nests and 4,718 ft [1,435 m] for unsuccessful nests). Prey deliveries averaged 3.0 fish per day (n=2 nests) and consisted primarily of sand lance (58 percent) and capelin (21 percent) with smaller amounts of smelt (9 percent), herring (6 percent) and snake prickleback (
Lumpenus sagitta;
6 percent). The mean number of days to fledging (±1 SD) at 9 nests was 23.7 (±3.5) days, or comparable to nests monitored at Kodiak Island.

In addition to Kittlitz's murrelets, researchers captured and radio-marked marbled murrelets in 2011 (n=7) and 2012 (n=9) in Icy Bay to compare reproductive performance between the two closely related species (Kissling, unpublished data). Across both years, 11 of 16 (69 percent) radio-marked marbled murrelets attempted to nest (two actually renested successfully), and 9 of 13 nests were successful (69 percent apparent nesting success). Marbled murrelet nests were located at lower elevations (median elevation=1,368 ft [417 m]) and closer to the ocean (median distance=2.9 mi [4.7 km]) than were Kittlitz's murrelet nests (4,291 ft [1,308 m] and 8.8 mi [14.2 km], respectively). Both breeding propensity and nest success of marbled murrelets were far greater than that for Kittlitz's murrelets using the same techniques in the same study area. Although the sample sizes are small, these results are important for two reasons: (1) It is unlikely that there was a capture, handling or radio-transmitter effect negatively biasing the poor reproductive measures of Kittlitz's murrelets, assuming that Kittlitz's and marbled murrelets would respond similarly; and (2) despite their similar life histories, Kittlitz's murrelets were consistently outperformed reproductively by marbled murrelets in Icy Bay, suggesting perhaps that forage-fish abundance was not limiting the nesting success of Kittlitz's murrelets. Possible reasons for the differences in reproduction of the two species are reduced foraging efficiency of Kittlitz's murrelets, availability of suitable nest sites, carry-over effects from the non-breeding period (Sorensen
et al.
2009, p. 464), or increased energetic costs of Kittlitz's murrelets to access nests at higher elevations and farther from the ocean (Hatch 2011, pp. 86-87).

Juveniles at sea
—Juvenile and adult Kittlitz's murrelets are readily distinguishable in hand owing to plumage characteristics, and usually, the presence of an egg-tooth in juveniles (Kissling, unpublished data); however, these identification markers are not easily observed at a distance at sea, especially in August when fledglings arrive on the water and adults begin their concurrent pre-basic molt (Kuletz
et al.
2008, p. 34). This complication may prevent the accurate estimation of juvenile abundance and ratios of juveniles to adults, both of which have been used as indices to annual reproductive success of marbled murrelets (Beissinger 1995, pp. 391-392; Kuletz and Kendall 1998, pp. 450-455; Beissinger and Peery 2007, pp. 297-298; Kuletz
et al.
2008, p. 85).

Day and Nigro (2004, pp. 91-93) suggested that reproductive success in Kittlitz's murrelets may be very low based on the near absence of juvenile birds in late summer surveys in Prince William Sound. In 3 combined years of at-sea surveys conducted between 15 July and 15 August in 1996, 1997, and 1998, in the fjords of Prince William Sound, only a single hatch-year bird was sighted (Day and Nigro 2004, p. 91). During similar late summer surveys in Kachemak Bay from 2004 to 2007, densities of juvenile Kittlitz's murrelets varied among years (range=0.01-0.05 birds per square mile (mi
2
) [0.03-0.12 birds per square kilometer (km
2
)]) and were much lower than those of marbled murrelets (range=0.10-0.31 birds per mi
2
[0.27-0.79 birds per km
2
]); however, juvenile to adult ratios were comparable between species ranging from 0.01 to 0.28 for Kittlitz's murrelets and from 0.02 to 0.13 for marbled murrelets, albeit with less intra-annual variation for the latter species (Kuletz
et al.
2008, pp. 59, 85). To provide perspective, the total number of juvenile Kittlitz's murrelets recorded in the 4 years of surveys was 37 among 1,445 sub-adults and adults (Kuletz
et al.
2008, pp. 104-107). Similarly, during surveys of nearshore waters around Kodiak Island in August 2011 and 2012, 16 juvenile and only 6 sub-adults and adult Kittlitz's murrelets were observed compared to 187 juvenile and 5,779 sub-adults and adult marbled murrelets (Corcoran 2012, p. 5). Between 2008 and 2011, only 5 juvenile to 380 adult Kittlitz's murrelets were captured in late summer in Icy Bay (Kissling, unpublished data). Thus, results of all of these studies are difficult to interpret without information on the behavior and timing of movements of both age classes of Kittlitz's murrelets in late summer and some estimates of detection errors. Fairly high ratios of juveniles to adults in Kachemak Bay and Kodiak Island suggest good reproductive performance in these areas, yet nest monitoring data on Kodiak Island indicate differently; therefore, the high ratios may reflect rapid and synchronous departure of adult Kittlitz's murrelets from these areas or post-fledging dispersal of juvenile Kittlitz's murrelets into these areas.

In Icy Bay, six juvenile Kittlitz's murrelets (1 immediately prior to fledging, 3 newly fledged, and 2 approximately 2-3 weeks post-fledgling) were captured and radio-marked in 2008-2010 (Kissling, unpublished data). All juveniles still had their egg-tooth at the time of capture. The 3 newly fledged birds were located within Icy Bay for approximately 24 hours before departing; 2 of them were not detected again, but 1 returned to Icy Bay 8 days later. The older fledglings, which were significantly heavier than the newly fledged birds, were relocated in Icy Bay for 1-3 weeks post-marking. All juveniles were relocated visually and appeared to be good swimmers and divers, although the newly fledged birds were not readily capable of flight, in contrast to the older fledglings that were excellent flyers and were indistinguishable from flying adults both in terms of flight ability and plumage. The small sample size precludes drawing definitive conclusions; however, these results suggest that most newly fledged Kittlitz's murrelets immediately depart their breeding area. After becoming proficient at foraging on their own, gaining weight and improving flight capability to avoid predators, they may return to their breeding area where they remain until the post-breeding migration begins. This possible scenario explains the differences in behavior between the newly fledged and post-fledged Kittlitz's murrelets. A better understanding of juvenile behavior after fledging would help to determine the reliability of juvenile surveys in late summer, which may be the most realistic and cost-efficient method for long-term monitoring of reproductive performance across many different study sites, as it is for marbled murrelets.

Survival.
The only estimates of survival of Kittlitz's murrelets were derived from data collected in Icy Bay. Using radio-marked Kittlitz's murrelets (n=197), Kissling (unpublished data) estimated breeding season survival (60 days post-marking; approximately mid-May through mid-July) of adults greater than 1 year old to be 0.89 (SE=0.04) with little inter-annual variation (n=6 years). The primary cause of adult mortality in the breeding season in Icy Bay was predation by peregrine falcons

(
Falco peregrinus
) and bald eagles (
Haliaeetus leucocephalus
). Based on mark-recapture banding methods, annual survival (1 July to 30 June) of adult Kittlitz's murrelets was estimated to be 0.80 (SE=0.33). Although this estimate is imprecise, primarily because of low recapture rates across years (less than 8 percent), it is comparable to annual survival of marbled murrelets (0.83-0.88) estimated using similar methods (Cam et al 2003, p. 1122; Peery
et al.
2006, p. 83). There are no estimates of juvenile survival of Kittlitz's murrelets, but estimates of annual survival of juvenile marbled murrelets range from 0.51 based on radio-telemetry (Parker
et al.
2003, p. 207) to a proportion of adult survival (70 percent) by comparing with other alcids (Nur 1993 in Piatt
et al.
2007, p. 55).

Population Status and Trends

In this section, we summarize information on status and trends of the Kittlitz's murrelet at the local population scale (i.e., by individual study areas) and at a broad scale across multiple populations. We also describe difficulties in estimating population size and trends of the Kittlitz's murrelet.

Estimating abundance and population trends for most alcids is simpler than for the Kittlitz's murrelet because the majority of alcids nest in colonies where birds concentrate and can be monitored in large numbers during the breeding season. In contrast, the solitary, remote, and secretive nesting behavior of the Kittlitz's murrelet makes terrestrial monitoring impractical for the purposes of estimating abundance and population trends (Drew and Piatt 2008, p. 179). Therefore, estimating abundance and the rate of change in populations of Kittlitz's murrelets has relied entirely on at-sea surveys (Day 2011, p. 2).

A handful of ornithological surveys and expeditions primarily aimed at documenting the distribution of marine birds occurred prior to 1972 (Isleib and Kessel 1973, p. 1), when systematic at-sea surveys were conducted in a few select locations in Alaska (Bailey 1977, p. 60; Klosiewski and Laing 1994, p. 5) and along discontinuous sections of shoreline in Russia (summarized in Artukhin
et al.
2011, pp. 25-26). Since then, many surveys for marine birds, including a number of efforts specifically for the Kittlitz's murrelet, covering a wider geographic area have been conducted and, in some areas, repeated in subsequent but not necessarily continuous years. These historical and recent survey efforts have provided a tremendous amount of information on the distribution and abundance of the Kittlitz's murrelet within the areas surveyed. Nonetheless, inherent, methodological, and analytical difficulties in estimating population size and trend of this species remain, many of which are not mutually exclusive and some of which can be resolved as new information becomes available.

First, present-day populations of Kittlitz's murrelet occupy a large range and are geographically clustered, usually in remote areas that are difficult to reach and survey. Many areas of their range have not yet been systematically surveyed or are under-represented by existing survey efforts.

Second, the high at-sea spatial and temporal variation of Kittlitz's murrelets often results in wide variances associated with population estimates and therefore little power to detect trend (Kissling
et al.
2007, p. 2168; Kirchhoff 2011, pp. 79-80; but see Drew
et al.
2008, pp. 18, 41). Each surveyed area differs in size, which has implications for estimating abundance. Surveys attempting to encompass larger areas, such as Prince William Sound, may encompass the spatial variability of murrelets that use this area during the summer; that is, surveys may be sufficiently large to encompass the spatial variation in areas used by murrelets during a survey effort owing to daily or weekly movements by murrelets within that area. However, larger areas take longer to survey and thus must capture the temporal variability in murrelet abundance. None of the survey areas, except Icy Bay (see
Nesting
and
Demography,
above), has been accompanied by related studies of daily (or longer) movements by murrelets to help understand whether the at-sea surveys are encompassing the range of habitats used by murrelets in that area during the survey period.

Third, the Kittlitz's murrelet can be difficult to distinguish from the more common marbled murrelet during surveys, resulting in varying proportions of
Brachyramphus
murrelets identified to genus only. This issue was particularly problematic during earlier surveys (pre-2000), when there was less emphasis and training on distinguishing between the two species during surveys of all marine birds, occasionally leading to high proportions (greater than 50 percent) of unidentified murrelets (Piatt
et al.
2011, p. 66; Day 2011, pp. 22-27; Kuletz
et al.
2011a, p. 99; Kuletz
et al.
2011b, pp. 87, 90) and possibly unknown proportions of misidentified murrelets (Kirchhoff 2011, pp. 80-81; Hodges and Kirchhoff 2012, p. 117; Kuletz
et al.
2013, p. 69).

Fourth, owing to their asynchronous arrival at breeding sites, unknown fidelity to breeding areas, and movements during the breeding season, it is difficult to define both a statistical or biological population of Kittlitz's murrelets. Therefore, apparent trend in local population size of the Kittlitz's murrelet is confounded by intra- and inter-annual movements of individuals among study sites. Most Kittlitz's murrelets apparently do not breed annually (Day and Nigro 2004, p. 91; Kissling, unpublished data) and, therefore, are not restricted to a particular breeding site or at-sea areas near a breeding site every year, allowing non-breeding individuals and failed breeders to move freely to locate food during the breeding season when most surveys are conducted. While breeding birds may not be counted on surveys because they are incubating or tending to young at nests, this probably is minor because breeding propensity typically appears to be low in this species (see
Reproductive Performance,
above).

In Icy Bay, the daily emigration rate of radio-marked Kittlitz's murrelets (±SE) over a 60-day period during the breeding season was low (0.008±0.002) (Kissling, unpublished data), but no estimate of the rate of immigration exists. Similarly, the annual recapture probability of uniquely banded Kittlitz's murrelets (±SE) was low (0.08±0.03), suggesting that individuals do return to the area, but perhaps not annually (Kissling, unpublished data). These intra- and inter-annual movements complicate reliable trend estimation of local population size, especially because the timing of at-sea surveys for the Kittlitz's murrelet has not been synchronized among study sites. To illustrate an extreme example, the local population of Kittlitz's murrelet in Kachemak Bay was estimated to be 1,776 birds (SE=1,051) in 2005, but 3,277 birds (SE=1,582) in 2006, followed by a drastic reduction in 2007 to 1,086 birds (SE=931) (Kuletz
et al.
2011b, p. 96). The documented fluctuations in local population size over the 3-year period cannot be demographically explained and therefore probably are related to intra- or inter-annual movements into or out of Kachemak Bay.

Fifth, there is not a consistent survey protocol or design used to count Kittlitz's murrelets at sea across locations and occasionally at the same location (Day 2011, pp. 6-39). Key survey and analytical procedures such as time of year and synchrony of counts across range, level of surveyor expertise and training, limitations of oceanic and climatic conditions, varying survey platforms, estimation of detection probabilities, inclusion of flying

murrelets, survey objectives (single-species versus multi-species surveys), and treatment of unidentified murrelets in population-size estimation have varied among locations and years (Hoekman
et al.
2011, p. 35; Kirchhoff 2011, p. 78; see Day 2011 for complete review). Within a study site, many methodological issues have been addressed in recent years, but across sites, inconsistencies will remain until a comprehensive monitoring protocol is developed, accepted and implemented by researchers. Until then, our ability to detect population trend of Kittlitz's murrelet reliably, especially beyond individual study sites, is limited (Day 2011, pp. 52-57).

Recognizing all of these challenges and differences in methods across study sites, the rangewide population of Kittlitz's murrelet currently is estimated to be 33,583 birds (95 percent CI=25,620-41,546). Because some areas remain unsurveyed or have not been surveyed in many years, this estimate should be considered a minimum. The rangewide estimate was derived by summing the most recent local population estimate in all surveyed areas during the breeding season, which includes all known concentrations of Kittlitz's murrelet. These areas (and most recent survey year) include the mainland fjords of southeastern Alaska (2002) (Kissling
et al.
2011, p. 7), Glacier Bay (2010-2012, averaged) (Hoekman
et al.
2013, p. 15), the outer coast of southeastern Alaska from Cross Sound to Yakutat (2003-2004) (Kissling
et al.
2011, p. 7), Yakutat Bay (2009) (Kissling
et al.
2011, p. 7), Lost Coast extending from Manby Point (2002) to Duktoth River (2008-2009) (Kissling
et al.
2011, p. 7), Icy Bay (2012) (Kissing, unpublished data), Kenai Fjords (2008) (Arimitsu
et al.
2011, p. 18), Prince William Sound (2012) (Cushing, Oregon State University, 2010-2012, unpublished data), Kachemak Bay (2011) (Kuletz, Service, 2011, unpublished data), Lower Cook Inlet (1996) (Kuletz
et al.
2011b, p. 96), Kodiak (2012) (Corcoran 2012, p. 5), southern coast of the Alaska Peninsula (2003) (Madison
et al.
2011, p. 118), select Aleutian Islands (2004-2009) (Madison
et al.
2011, p. 118), northern Alaska including Chukchi and Beaufort seas and Arctic Ocean (2000-2009) (Day
et al.
2011, p. 58), eastern coast of Russia extending from the Chukotka Peninsula in the north to the southern tip of the Kamchatka Peninsula (1991-2005) (Artukhin
et al.
2011, pp. 26-28) and the northern Sea of Okhotsk (2005-2008) (Artukhin
et al.
2011, p. 30). Hence, this population estimate does not include numbers from Kodiak Island, most of the Aleutian Islands, and the Bering Sea, plus non-breeding birds that may be living at sea across the open northern Gulf of Alaska.

We examined trends of the Kittlitz's murrelet at the local population scale and across multiple populations. We defined a population as the birds using pre-defined study area boundaries, although there is no evidence that these individual populations are biologically separated from one another. Below, we briefly summarize available information about local population trends of the Kittlitz's murrelet in areas for which a sufficient number of years of data were available. In many cases, we were unable to draw inferences on the trend of Kittlitz's murrelet at the local population scale because of the difficulties described above and, in some cases, conflicting information within a study area. However, we did not consider this to be a limitation to our assessment because our primary interest was to determine, to the best of our ability, the status and trend of the Kittlitz's murrelet at a broad scale, as opposed to the local population scale. Therefore, we analyzed trend across multiple populations of Kittlitz's murrelet using all of the available information on local populations with at least 3 years of at-sea survey data and developed a population model that also incorporated information on reproduction and survival; these two efforts to assess the status and trend of Kittlitz's murrelet across multiple populations are also summarized below.

Local Populations

Only 7 areas have been surveyed for Kittlitz's murrelets at sea in a somewhat consistent manner in 3 or more different years between 1989 and 2012: Glacier Bay (13 surveys of 3 different designs between 1991 and 2012), Malaspina Forelands (4 surveys of one continuous transect, 1992-2009), Icy Bay (2002-2012), Prince William Sound (13 years, 1972 and 1989-2012, with a different design in 1972), Kenai Fjords (3 years, 2006-2008), Kachemak Bay (4 years, 2005-2011), and Lower Cook Inlet (5 years, 1993-1999 using two different designs). Few surveys were conducted prior to 2000, and the reliability of some of those survey data is compromised due to the methodological challenges presented above. Therefore, rates of change in local population size in the few areas where early surveys were completed (i.e., Glacier Bay, Malaspina Forelands, Prince William Sound, and Lower Cook Inlet) often rely heavily on 1-2 historical years of data.

Glacier Bay (37 percent of rangewide population estimate).
Three different research teams have conducted systematic at-sea surveys for marine birds in Glacier Bay and all have employed their own survey design and protocol (Kirchhoff 2011, p. 78). Piatt
et al.
(2011, p.71) conducted surveys in 1991, 1999-2003, and 2008 and reported a local population decline of 89 percent (negative 10.7 percent per year) over this time period, but the decline was not statistically significant due to high inter- and intra-annual variance. During the 1991 surveys, a different sampling design was used that did not sample the offshore habitat randomly nor systematically, raising concern that the 1991 survey results were not comparable to data collected in 1999-2008 (Drew and Piatt 2008, p. 179; Day 2011, p. 39; Kirchhoff 2011, p. 78). However, the authors believed that they had adequately addressed discrepancies between the two designs in their analysis and that Kittlitz's murrelets, in fact, did decline in Glacier Bay between 1991 and 2008, although the decline appeared to level off after 2002 (Piatt
et al.
2011, p. 72). Further, Drew and Piatt (2008, p. 178) tested for potential survey-based bias in both sampling designs by using spatially matched transects and concluded that the Kittlitz's murrelet population in Glacier Bay had declined by 83 percent between 1991 and 2000. During a similar time period, Lindell (2005, p. 5) conducted surveys in 1993 in Glacier Bay that were replicated in 2009 and 2010 (Kirchhoff
et al.
2013, p. 6). When analyzed collectively with survey results completed by Piatt
et al.
(2011, p. 7), the annual rate of change was negative 2.3 percent between 1991 and 2010 and was not statistically significant (Kirchhoff
et al.
2013, p. 10). Most recently, Hoekman
et al.
(2011, p. 35; 2013, p. 15) developed and tested a new, sophisticated survey design and protocol specifically for Kittlitz's murrelets in Glacier Bay and completed annual surveys accordingly in 2010-2012; this protocol currently is under review. The field and analytical techniques employed by Hoekman
et al.
(2013, p. 15) have resulted in much larger population estimates (2-3 times greater) of Kittlitz's murrelet compared to the more standard approaches used by other researchers (Lindell 2005, p. 5; Piatt
et al.
2011, p. 71; Kirchhoff
et al.
2013, p. 6). There is notable disagreement among researchers about the current population size and trend of Kittlitz's murrelets in Glacier Bay, with the disagreement about trends primarily

due to differences in survey design and protocol of the 1991 survey.

Malaspina Forelands (less than 1 percent of rangewide population estimate).
The Malaspina Forelands, an area extending between Manby Point near Yakutat and Point Riou at the entrance to Icy Bay, was surveyed initially in 1992 by paralleling the coastline roughly 0.6 mi (1 km) offshore for 51 mi (82 km) (Kozie 1993, pp. 1-2). Kissling
et al.
(2011, p. 4) repeated this survey in 2002, 2008, and 2009. Results of these surveys are useful to document distribution and qualitative rates of change, but the survey design of one linear transect lacks rigor and does not lend itself to estimation of population size or trend at the scale of a study area. In the four surveys, however, the number of Kittlitz's murrelets varied dramatically ranging from 641 in 1992 to 10 (2002), 39 (2008) and 165 (2009) (Kissling
et al.
2011, p. 7).

Icy Bay (3 percent of rangewide population estimate).
Since 2002, eight at-sea surveys for marine birds targeting Kittlitz's murrelet and using the same study design and sampling methods have been conducted in Icy Bay (2002, 2005, 2007-2012; Kissling
et al.
2011, p. 7; Kissling, unpublished data). Between 2002 and 2012, the annual rate of change of the local population was estimated to be negative 10.0 percent; this rate of decline was statistically significant (slope estimate=negative 0.10 [SE=0.03]) (Kissling, unpublished data). A population model that incorporates demographic characteristics including reproduction, survival, and abundance of Kittlitz's murrelets in Icy Bay substantiated the results of the at-sea surveys by indicating an 8 percent decline annually between 2002 and 2012, but the variance surrounding this estimate is large (Kissling, unpublished data).

Prince William Sound (4 percent of rangewide population estimate).
While Prince William Sound has the longest history of survey effort (13 years), it is also subject to reliability concerns related to historical data, especially in regards to varying proportions of unidentified
Brachyramphus
murrelets, as well as impacts of the
Exxon Valdez
oil spill that occurred in March 1989. Several authors determined that there had been declines in some marine bird populations, including
Brachyramphus
murrelets, before the spill occurred (Klosiewski and Laing 1994, p. 28; Agler
et al.
1999, p. 101). Kuletz
et al.
(2011a, p. 103) reported a decline in Kittlitz's murrelets in Prince William Sound of 18.1 percent per year between 1972 and 2007. However, interpretation of population trend in this area was complicated by three primary concerns: (1) The 1972 survey used a different survey design than the 1989-2012 surveys; (2) the 1972 survey was temporally removed from the remainder of the surveys (17 years between the 1972 survey and the next survey in 1989); and (3) the earlier surveys in 1989-1991 and 1993 had high proportions of unidentified murrelets (39-89 percent). Kuletz
et al.
(2011a, pp. 99-101) developed a population model that included the unidentified murrelets to overcome these challenges and tested the sensitivity of the model to inclusion and exclusion of the problematic years; after acknowledging the many assumptions and limitations of the analysis, the authors found negative trends for populations of Kittlitz's murrelets in Prince William Sound regardless of which years were included (p. 104). In fact, even after removing the 1972 survey, the annual rate of population change of Kittlitz's murrelets was greater (negative 30.0 percent; Kuletz
et al.
2011a, p. 103).

Hodges and Kirchhoff (2012, pp. 118-119), however, postulated that misidentification of murrelets in 1989 and 1993 was probable. Based on a reanalysis excluding those years and including 2 additional survey years (2001 and 2009; not included by Kuletz
et al.
(2011a, p. 101) because only select fjords within Prince William Sound representing different statistical populations of murrelets were surveyed in these years), Hodges and Kirchhoff (2012, p. 119) concluded that population trend of Kittlitz's murrelet in Prince William Sound between 1989 and 2009 was not significantly different from a stable population. Kuletz
et al.
(2013, pp. 69-71) disputed the case presented by Hodges and Kirchhoff (2012, pp. 118-119), contending that the authors erred in their assumption of Kittlitz's murrelet distribution, and in including the 2001 and 2009 survey data, concluding that there was in fact a decline in the Kittlitz's murrelet in Prince William Sound between 1989 and 2007.

Cushing
et al.
(2013, p. 1) took a different approach to address the high and varying proportions of unidentified and possibly misidentified murrelets by simply reporting populations trends of
Brachyramphus
murrelets (genus level) in Prince William Sound between 1989 and 2012. There was strong evidence of an overall decline in abundance of murrelets with a mean annual rate of change of negative 5.2 percent or a 70.8 percent cumulative decrease in abundance of
Brachyramphus
murrelets over the 23-year period. This estimate of decline applies to both Kittlitz's and marbled murrelets, however, so it is difficult to draw firm conclusions about the status of Kittlitz's murrelets in Prince William Sound from this analysis. Kuletz
et al.
(2013, pp. 69-71) argued that, given the undisputed decline in
Brachyramphus
murrelets in Prince William Sound, the proportion of identified Kittlitz's murrelets to marbled murrelets should have increased if the Kittlitz's murrelet population was stable, but instead the proportion of identified Kittlitz's murrelets has declined between 1989 and 2012. However, this argument hinges on comparable identification rates of both murrelet species within and among years.

Kenai Fjords (2 percent of rangewide population estimate).
Seven surveys using five different survey designs or protocols have been conducted in Kenai Fjords, prohibiting reliable estimation of local population trends of Kittlitz's murrelet. Arimitsu
et al.
(2011, p. 17) summarized earlier survey efforts for marine birds in the greater Kenai Fjords area (1976, 1986, 1989, 2002), most of which concentrated survey effort along the shoreline and did not follow a consistent survey protocol with previous surveys. Acknowledging many methodological issues associated with these surveys, density estimates of Kittlitz's murrelet increased by 55 percent between 1986 and 1989 and decreased by 90 percent between 1989 and 2002 (Arimitsu
et al.
2011, p. 18). In 2006-2008, annual surveys for Kittlitz's murrelets were conducted while following a systematic study design and sampling protocol similar to those employed in other areas, but with only 3 years of data over a short time frame, Arimistu
et al.
(2011, p. 17) appropriately refrained from estimating local population trend and instead assessed variability of the Kittlitz's murrelet population during the 3-year period.

Lower Cook Inlet and Kachemak Bay (9 percent of rangewide population estimate).
Of all areas with multiple years of surveys, Lower Cook Inlet and adjacent Kachemak Bay in the southeastern part of the inlet, are the most complex and confounding. In June 1993, Agler
et al.
(1998, pp. 255-256) completed a comprehensive, systematic survey for marine birds and mammals covering all of Lower Cook Inlet. A portion (roughly one-third) of this area was surveyed in July and August 1996-1999, but while using a different systematic sampling design (described in Kuletz
et al.
2011b, p. 86). Kuletz
et al.
(2011b, p. 86) reanalyzed data from a `core area' of Cook Inlet that had been

covered during both earlier survey efforts. Within the core area, numbers of Kittlitz's murrelet declined significantly by 26.2 percent per annum between 1993 and 1999, a total decline of 84 percent over the 7-year period (Kuletz
et al.
2011b, p. 91); however, there are two primary concerns related to these surveys. First, in 1993, 82 percent of the murrelets observed were not identified to species (Kuletz
et al.
2011b, p. 91), and second, the timing of the four surveys varied dramatically, especially between the 1993 survey (7-23 June) and the 1996-1999 surveys (14 July-16 August; p. 87), severely reducing the comparability of these surveys across years. Removing the 1993 survey from the trend analysis, numbers of Kittlitz's murrelet declined by 32 percent annually between 1996 and 1999 in the core area (Kuletz
et al.
2011b, p. 91), although these surveys started and ended later each consecutive year (p. 87). These results may be questionable, however, given recent information that Kittlitz's murrelets from other parts of the northern Gulf of Alaska are known to move into Lower Cook Inlet in the post-breeding season (late July-August; Madison
et al.
2012, p. 1).

Similarly, several late-summer surveys of varying sampling designs and protocols were conducted between 1988 and 2011 in Kachemak Bay (Kuletz
et al.
2011b, p. 90; Kuletz, unpublished data), but many of these survey efforts lacked a rigorous or systematic survey design, and there are concerns about the timing of the surveys. Therefore it is difficult to draw statistical inference from their results. Between 2005 and 2007, systematic surveys of Kachemak Bay were conducted from 18 to 26 July using standard protocols (Kuletz
et al.
2011b, p. 90), resulting in annual local population estimates ranging from 1068 to 3287 Kittlitz's murrelets, depending on the year. Based on these surveys, as well as the historical efforts, Kuletz
et al.
(2011b, p. 93) concluded that the population of Kittlitz's murrelet in Kachemak Bay was statistically stable. In 2011, the same systematic survey of Kachemak Bay was repeated, resulting in a considerably lower estimated population size of Kittlitz's murrelet (424 birds) than the previous 3 surveys completed in 2005-2007 (Kuletz, unpublished data). However, we cannot draw reliable conclusions from these data for two reasons. First, the variance associated with these local population estimates is too high to detect a trend between 2005 and 2011 (coefficient of variation [a measure of variability in the data]=52-86 percent) (Kuletz
et al.
2011b, p. 96; Kuletz, unpublished data). Second, the rate of change in population size was not linear across the 6-year period and the range in estimates (424 to 3,287 Kittlitz's murrelets) cannot be demographically explained (Kuletz
et al.
2011b, p. 96; Kuletz, unpublished data).

Multiple Populations

Trend analysis.
We assessed change in Kittlitz's murrelet populations at a broad scale by conducting a comprehensive trend analysis that used survey data collected at multiple individual study sites (hereafter referred to as the multiple-populations trend analysis) (Lukacs and Kissling 2013, p. 27). We limited our analysis to those areas with at least 3 different years of survey data, and within a study area, we only grouped surveys that sampled similar statistical populations; no datasets were combined. We included 9 statistically-independent populations with datasets spanning from 1989 to 2012 in the multiple-populations trend analysis: Glacier Bay-A (1991, 1999-2003, 2008) (Piatt
et al.
2011, p. 70), Glacier Bay-B (1993, 2009-2010) (Lindell 2005, p. 5; Kirchhoff
et al.
2012, pp. 6, 10), Glacier Bay-C (2010-2012) (Hoekman
et al.
2013, p. 15), Malaspina Forelands (1992, 2002, 2008-2009) (Kissling
et al.
2011, p. 7), Icy Bay (2002, 2005, 2007-2012) (Kissling
et al.
2011, p. 7; Kissling, unpublished data), Prince William Sound (1989-1991, 1993, 1996, 1998, 2000, 2004-2005, 2007, 2010, 2012) (Cushing, unpublished data), Kenai Fjords (2006, 2007, 2008) (Arimitsu
et al.
2011, p. 18), Kachemak Bay (2005-2007, 2011) (Kuletz
et al.
2011b, p. 96; Kuletz, unpublished data), and Lower Cook Inlet (1993, 1996-1999) (Kuletz
et al.
2011b, p. 96).

We considered four model forms to describe and estimate population trend of Kittlitz's murrelets across multiple local populations between 1989 and 2012: constant (no change over time), linear (straight line), quadratic (line that displays concavity with a single bend either upward or downward), and linear with a change in slope (statistically referred to as a `knot') at 2000 (Lukacs and Kissling, p.27). We tested the last model form (linear with a knot at 2000) because around this time climate regime shifts occurred in the northern Gulf of Alaska (1998-1999) and in the Arctic (2000) (Litzow 2006, p. 1386; Overland
et al.
2008, p. 92) (see
Factor A
below for more detailed discussion on climate regime shifts) and researchers reported that Kittlitz's murrelet numbers may have stabilized in some areas shortly thereafter (Kuletz
et al.
2011a, p. 105; Piatt
et al.
2011, p. 73). Of the four model forms considered in the multiple-populations trend analysis, the linear model form with a knot at 2000 was the most strongly supported model (delta Akaike Information Criterion [AIC]=19.2 units; AIC is a measure of the relative quality of a statistical model for a given set of data and contending model forms; a small delta AIC [e.g., less than 2] indicates model uncertainty).

Results of the multiple-populations trend analysis demonstrated that the population of Kittlitz's murrelet declined significantly by 30.6 percent per annum between 1989 and 2000 (slope estimate=negative 0.31 [SE=0.09]), at which time a statistically significant change in the rate of change occurred and populations stabilized between 2000 and 2012 (slope estimate=0.38 [SE=0.13]; this slope estimate represents the positive change from negative 0.31, or a positive slope of 0.07). We then removed 3 problematic years of data due to high proportions of unidentified murrelets (1993 in Prince William Sound and 1993 in Lower Cook Inlet) and to differences in study design (1991 in Glacier Bay) and reran the analysis. Although model fit with the problematic data points removed gave a poorer fit (delta AIC=12.2), the same model (linear with a knot at 2000) was selected and estimated similar trends across all populations between 1989 and 2000 (slope estimate=negative 0.30 [SE=0.10] and between 2000 and 2012 (slope estimate=0.38 [SE=0.14]. We conclude from this analysis that Kittlitz's murrelets declined by roughly 30 percent per annum on average across multiple populations between 1989 and 2000, after which abundance stabilized. For comparison, the same analysis for the population of marbled murrelet across multiple populations indicated a stable trend from 1989 to 2012 with no change in slope at year 2000; the constant model for marbled murrelet was selected as the best model (delta AIC=3.3) when we ran the analysis with and without the three questionable data points.

For assessing status of the Kittlitz's murrelet across their range, we found that the multiple-populations trend analysis described above is more useful and rigorous than trend estimates of individual local populations; however, several drawbacks to our approach exist. First, the trend analysis included populations of Kittlitz's murrelets only from Glacier Bay in the south to Lower Cook Inlet in the north, an area that contains most of the known larger populations of the Kittlitz's murrelet (see
Local Populations,
above), but covers a small portion of their overall

range. Second, we only considered linear and quadratic shapes to the trend of multiple populations combined. Third, demographic parameters such as reproduction and survival are not considered in the trend analysis, even though these vital rates drive current and future abundance. Fourth, the trend analysis does not allow population projections into the future or estimation of extinction probabilities. To address some of these drawbacks, we developed a population model as a tool to assessing population status of the Kittlitz's murrelet at a broad scale.

Population model.
Owing to the limitations of the multiple-populations trend analysis, we developed a population model to help evaluate the status of the Kittlitz's murrelet across all populations with sufficient demographic information (hereafter referred to as the multiple-populations model) (Brooks
et al.
2004, p. 515; Johnson
et al.
2010, p. 1084; Lukacs and Kissling 2013, p. 5). Population models are a well-established tool for evaluating population dynamics for species with limited and variable datasets, such as the Kittlitz's murrelet, by linking population size with stage-specific vital rates. A single comprehensive population model like the one we developed integrates all of the available data on abundance, survival, and reproduction; shares information from data-rich areas with data-poor areas; and predicts population size given the demographic data each year and into the future. One advantage to using this approach for the Kittlitz's murrelet is that it allowed us to include data on reproduction at Agattu and Kodiak islands and Icy Bay (see
Nesting
and
Reproductive Performance,
above) and on survival from Icy Bay (see
Survival,
above), thereby nearly doubling the spatial scope of inference compared to that of the multiple-populations trend analysis. Another advantage is that it is not purely a statistical test such as the multiple-populations trend analysis, but instead incorporates aspects of the biology of the Kittlitz's murrelet.

We included 7 local populations in the multiple-populations model: Glacier Bay, Icy Bay, Prince William Sound, Kenai Fjords, Kachemak Bay, Kodiak Island, and Agattu Island. In Glacier Bay, where multiple datasets on abundance exist, we used the dataset (Glacier Bay-A) with the most number of years of abundance estimates (Piatt
et al.
2011, p. 70). We modeled data collected from 2000 to 2012 because only abundance was available prior to 2000, and without concurrent data on reproduction or survival, we were unable to achieve a good model fit (Lukacs and Kissling 2013, p. 6). Because our primary interest was to determine the current and future status and population dynamics of the Kittlitz's murrelet at a broad scale and few demographic data were collected prior to 2000, we did not consider the exclusion of pre-2000 data to be a major constraint to the model development or results. Reproduction was estimated as the product of breeding propensity (the proportion of birds attempting to nest in a given year) and nesting success. Following Peery and Henry (2010, p. 2417), we considered a range of values for breeding propensity (low=0.181, medium=0.526, high=0.817; see
Reproductive Performance,
above, for details) (Kissling, unpublished data) and estimated daily nest survival at Agattu and Kodiak islands (0.968) (Kaler, unpublished data; Lawonn, Oregon State University, 2008-2011, unpublished data) and Icy Bay (0.979) (Kissling, unpublished data). For areas without nesting information, we applied the estimate of nesting success from the study site most similar in landscape (e.g., glacial, non-glacial). We considered a range of values for annual adult survival (low=0.79, medium=0.89, high=0.95) (Kissling, unpublished data) and used a proportion (0.70) of adult survival as juvenile survival following Peery and Henry (2010, p. 2415) and others (McShane
et al.
2004, p. 3-41; Piatt
et al.
2007, p. 58).

The best-fit model for the multiple-populations model included the medium-level breeding propensity (0.526) and medium-level annual survival (0.89) and predicted an annual rate of change in multiple populations to be negative 1.7 percent but with large variance that included both a stable population and a quasi-extinction scenario (Lukacs and Kissling 2013, p. 10). The probability of extinction, with a quasi-extinction threshold defined for the purposes of this modeling exercise as less than 100 individuals per population, at 2032 (i.e., 20 years from present) was zero and at 2037 (i.e., 25 years from present) was less than 0.01 (Lukacs and Kissling 2013, p. 10; Lukacs, University of Montana, unpublished data). We were unable to model population size accurately or precisely beyond 25 years into the future because the variance increased rapidly and the model became unstable. Given the paucity of data available for the Kittlitz's murrelet, predicting future population size is challenging for any number of years and becomes more difficult with increased time, but after examining model fit and diagnostics, we determined that model predictions of population size of this species between 2000 and 2037 were informative in our assessment of the current and future status of this species.

As with all modeling exercises, there are numerous limitations and assumptions related to model structure and inputs that need to be met or evaluated to assess reliability and usefulness of the model results. Key assumptions for this type of modeling (not a comprehensive list) include: (1) The model structure accurately represented Kittlitz's murrelet population biology; (2) populations were sampled independently; (3) populations are not under density-dependent regulation; (4) estimates of reproduction and survival were appropriately applied to and representative of populations lacking those data; (5) the populations for which sufficient data exist to include in the model were representative of all Kittlitz's murrelet populations; (6) immigration and emigration rates within a population were equal; and (7) estimates of vital rates and their associated variances between 2000 and 2012 that were used in the model to predict future population size will be comparable on average to those experienced by Kittlitz's murrelets between the present time and 2037. These are reasonable assumptions to make for the purposes of this modeling exercise in the absence of more complete data on the Kittlitz's murrelet or a similar species that would allow explicit testing of each assumption.

We acknowledge that the available information on the demography of the Kittlitz's murrelet is both spatially and temporally limited and therefore, attempted to account for these data limitations in the multiple-populations model in three ways. First, we chose to use a type of model (Bayesian Integrated Population Model) that is specifically aimed to serve as a powerful statistical tool for evaluating the dynamics of populations with messy or incomplete datasets (Brooks
et al.
2004, p. 515; Johnson
et al.
2010, p. 1084). Second, we considered a range of values for key demographic parameters such as breeding propensity and adult survival, placing weight on empirical data derived from that population and reducing weight for data borrowed from a different population. This approach allowed the empirical data available for a specific population to have a strong influence on the model results for that population. Third, we drew on previous population modeling efforts for the congeneric marbled murrelet, recognizing that all of these efforts, including our effort for the Kittlitz's

murrelet, had different objectives and therefore used a different type of population model (Beissinger 1995, pp. 385-393; McShane
et al.
2004, pp. 3-27-3-58; Piatt
et al.
2007, pp. 54-67; Peery and Henry 2010, pp. 2414-2424). We also used the marbled murrelet as a proxy for some unknown or less-defined demographic parameters of the Kittlitz's murrelet. We recognize all of these limitations and assumptions of the multiple-populations model and believe that the high variance associated with most of the model input parameters and the results accurately reflects our current state of knowledge of the status of the Kittlitz's murrelet at a broad scale.

Summary of Population Status and Trends

We estimate the minimum rangewide population of Kittlitz's murrelet to be 33,583 birds (95 percent CI=25,620-41,546). In evaluating population status and trends of the Kittlitz's murrelet, we collectively considered all of the available information across all time periods, at the local population scale, and at a broad scale across multiple populations. We determined that some local populations of the Kittlitz's murrelet may have declined at some point over the last few decades (e.g., Glacier Bay, Prince William Sound, Lower Cook Inlet) and some may still be in decline (e.g., Icy Bay, Kachemak Bay). Across all populations, we conclude that there was a decline of approximately 30 percent per annum in Kittlitz's murrelets between 1989 and 2000, but since then populations appear to have stabilized or, when coupled with information on reproduction and survival, may be declining and are projected to continue to decline at a much slower rate.

Summary of Information Pertaining to the Five Factors

Section 4 of the Act (16 U.S.C. 1533) and implementing regulations (50 CFR 424) set forth procedures for adding species to, removing species from, or reclassifying species on the Federal Lists of Endangered and Threatened Wildlife and Plants. Under section 4(a)(1) of the Act, a species may be determined to be endangered or threatened based on any of the following five factors:

(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 pertaining to the Kittlitz's murrelet in relation to the five factors provided in section 4(a)(1) of the Act is discussed below. In considering what factors might constitute threats, we must look beyond the mere exposure of the species to the factor to determine whether the species responds to the factor in a way that causes actual impacts to the species. If there is exposure to a factor, but no response, or only a positive response, that factor is not a threat. If there is exposure and the species responds negatively, the factor may be a threat and we then attempt to determine how significant a threat it is. If the threat is significant, it may drive or contribute to the risk of extinction of the species such that the species warrants listing as endangered or threatened as those terms are defined by the Act. This does not necessarily require empirical proof of a threat. The combination of exposure and some corroborating evidence of how the species is likely impacted could suffice. The mere identification of factors that could impact a species negatively is not sufficient to compel a finding that listing is appropriate; we require evidence that these factors are operative threats that act on the species to the point that the species meets the definition of an endangered or threatened species under the Act.

In making our 12-month finding on the petition we considered and evaluated the best available scientific and commercial information.

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

The Kittlitz's murrelet is primarily a subarctic species with a broad distribution that encompasses a diversity of marine and terrestrial habitats along most of coastal Alaska and eastern Russia. This species exhibits variable habitat affinities throughout its range and its annual cycle, which makes it difficult to identify necessary habitats and potential threats to those habitats. For example, in the breeding season, the greatest densities of Kittlitz's murrelet typically are observed in glacially-influenced marine waters of south-coastal Alaska (Kuletz
et al.
2003b, p. 136; Arimitsu
et al.
2011, p. 18; Kissling
et al.
2011, p. 7; Kuletz
et al.
2011a, pp. 102-103; Kuletz
et al.
2011b, pp. 90-92; Piatt
et al.
2011, p. 70). There are a handful of glaciated areas in southeastern Alaska, such as Le Conte, Thomas, Dundas, and Taylor bays, where no Kittlitz's murrelets have been observed in the breeding season in recent years, although several individuals were collected historically in Le Conte Bay (Kissling
et al.
2011, pp. 7, 9). Lower densities of this species also occur in non-glaciated marine waters of the Alaska Peninsula, Aleutian Islands (Madison
et al.
2011, pp. 118-119), western and northern Alaska (Day
et al.
2011, pp. 58-59) and Russia (Artukhin
et al.
2011, pp. 26-30). Low numbers of Kittlitz's murrelet also have been observed annually during the breeding season on freshwater lakes in southwestern Alaska (Savage 2013, in litt.; Walsh 2013, in litt.). In the non-breeding season, Kittlitz's murrelets migrate to the Bering and Chukchi seas where they occupy offshore marine waters, or occur in polynyas or in open water leads within the sea ice (Madison
et al.
2012, p. 1; Kuletz, unpublished data), but they also are observed in ice-free waters of the northern Gulf of Alaska during this period (Day
et al.
1999, pp. 4-5; Kuletz, unpublished data). The reason for the apparent, but irregular, association with sea ice or glacial ice during specific periods in the annual cycle is not clear, nor is it known if it is biologically meaningful or is simply a proxy for an unidentified habitat feature of importance (Arimitsu
et al.
2012, p. 18). Furthermore, it is not known whether the explanatory factor(s) occur in the marine or terrestrial habitat of the Kittlitz's murrelet, or both.

Without an understanding of the habitat requirements of the Kittlitz's murrelet, we identified, deconstructed, and assessed possible threats to the marine and terrestrial habitats currently used by this species. We then evaluated potential impacts by considering the exposure and response of Kittlitz's murrelet at the individual level and population level to each possible threat. Because the underlying mechanisms driving habitat use of the Kittlitz's murrelet are not defined, we attempted to establish links between possible threats to marine and terrestrial habitats and demographic change of Kittlitz's murrelet at the population level. Our analysis focused on possible threats to habitats occupied by Kittlitz's murrelets in the summer months because this was the time period for which the most data were available, along with the greatest number of possible identified threats and demographic bottlenecks (e.g., poor reproduction; see
Reproductive Performance,
above). We considered potential threats during the non-breeding period if sufficient information was available. Under
Factor A,
we considered climate change and

environmental contaminants as potential threats to the habitats used by the Kittlitz's murrelet.

Climate Change

Our analyses under the Act include consideration of ongoing and projected changes in climate. The terms “climate” and “climate change” are defined by the Intergovernmental Panel on Climate Change (IPCC). “Climate” refers to the mean and variability of different types of weather conditions over time, with 30 years being a typical period for such measurements, although shorter or longer periods also may be used (IPCC 2007, p. 78). The term “climate change” thus refers to a change in the mean or variability of one or more measures of climate (e.g., temperature or precipitation) that persists for an extended period, typically decades or longer, whether the change is due to natural variability, human activity, or both (IPCC 2007, p. 78). Various types of changes in climate can have direct or indirect effects on species. These effects may be positive, neutral, or negative, and they may change over time, depending on the species and other relevant considerations, such as the effects of interactions of climate with other variables (e.g., habitat fragmentation) (IPCC 2007, pp. 8-14, 18-19). Identifying likely effects often involves aspects of climate change vulnerability analysis. Vulnerability refers to the degree to which a species (or system) is susceptible to, and unable to cope with, adverse effects of climate change, including climate variability and extremes. Vulnerability is a function of the type, magnitude, and rate of climate change and variation to which a species is exposed, its sensitivity, and its adaptive capacity (IPCC 2007, p. 89; see also Glick
et al.
2011, pp. 19-22). There is no single method for conducting such analyses that applies to all situations (Glick
et al.
2011, p. 3). We use our expert judgment and appropriate analytical approaches to weigh relevant information, including uncertainty, in our consideration of various aspects of climate change.

Within the range of the Kittlitz's murrelet, climate change is occurring and is likely altering marine and terrestrial habitats used by this species. In Alaska, average annual Statewide air temperatures have increased by nearly 4.0 degrees F between 1949 and 2005, but decreased by 2.3 degrees F in the last decade (2000-2010) with most of the change occurring over winter (Markon
et al.
2012, p. 11; Wendler
et al.
2012, pp. 111-112). The recent cooling trend suggests a shift from the long-term warming trend (Wendler
et al.
2012, p. 111), even though climate models project warming to continue in Alaska over the next century (Markon
et al.
2012, pp. 14-21). Precipitation also increased over the last few decades, but it is more difficult to quantify (Arendt
et al.
2009, p. 4132; Markon
et al.
2012, p. 12). In addition, subsurface and surface waters of the North Pacific Ocean, including the Gulf of Alaska, and Bering and Chukchi seas, have warmed over the last few decades (Bograd
et al.
2005, p. 244; Overland and Wang 2007, p. 178; Stabeno
et al.
2007, pp. 2607-2608; Steele
et al.
2008, p. 2; Mueter
et al.
2009, p. 96; Hazen
et al.
2012, p. 2). A significant part of the observed warming in Alaska occurred as a sudden, step-like change in the mid-1970s, which coincided with a major shift in atmospheric circulation patterns across a large portion of the Pacific basin, called the Pacific Decadal Oscillation (PDO) (Mantua
et al.
1997, p. 1070). It is likely that some portion of the observed warming over the last century and recent cooling in Alaska is attributed to inherent decadal-scale variability in regional climate, like the PDO (Markon
et al.
2012, p. 11; Wendler
et al.
2012, p. 113), making it difficult to ascertain any amplified or accelerated impacts of natural variability or cycles from underlying long-term warming trends in Alaska. Regardless, marine and terrestrial habitats of the Kittlitz's murrelet are changing in response to climate change, and we anticipate that these changes will continue. Available information suggests that the changes may affect the Kittlitz's murrelet; however, the specific response or sensitivity of the species to these current and forecasted changes is uncertain at this time.

Loss of Glaciers

Loss of glacial volume is a phenomenon occurring on a global scale and, during the recent decades, at rates that cannot be explained by historical trends alone (Dyurgerov and Meier 2000, pp. 1406, 1410; Lemke
et al.
2007, pp. 356-359). The primary driver of glacier change is climate (Markon
et al.
2012, p. 45), especially temperature (Oerlemans 2005, p. 677; Arendt
et al.
2009, p. 4132). Maritime glaciers terminating in tidewater are particularly sensitive to temperature change (Berthier
et al.
2010, p. 93), including sea surface temperatures (Post
et al.
2011, p. 306), and therefore have the potential to shed ice more rapidly than land-locked glaciers (Markon
et al.
2012, p. 46). Yet, changes in individual tidewater glaciers are dominated by dynamic, complex cycles, with low-order effects occurring due to climate (Arendt
et al.
2009, p. 4132; Post
et al.
2011, p. 306).

At the beginning of the 20th century, many of Alaska's tidewater glaciers began to retreat (Barclay
et al.
2006, p. 160) and in less than 100 years, major ocean inlets, such as Glacier and Icy bays, were formed by glacial recession (Molnia 2008, p. K7). The Kittlitz's murrelet presumably adjusted its distribution in order to take advantage of these newly-created habitats where they now occur in large numbers in the breeding season (Kissling
et al.
2011, p. 7; Piatt
et al.
2011, p. 66). Currently, within the range of the Kittlitz's murrelet, 59 major tidewater glaciers exist, all along the southern coast of Alaska (Molnia 2008, pp. K57-59), and a few very small isolated mountain glaciers or permanent snow occur on the Alaska Peninsula, select Aleutian Islands, Koryak Highlands and Kamchatka Peninsula (Artukhin et al 2011, p. 31; Arendt
et al.
2012). The majority (68 percent) of these tidewater glaciers are in retreat, grounded (resting on the ocean floor) or at the shoal (shallow water area) (Molnia 2008, pp. K57-59). Over the last few decades, glacial ice loss has been greatest for the glaciers along the southern coast of Alaska compared to the mountain glaciers of central Alaska, Brooks Range, and Alaska Peninsula (Larsen
et al.
2005, p. 548; Berthier
et al.
2010, pp. 92-93; Arendt
et al.
2009, pp. 4127-4128; Le Bris
et al.
2011, p. 141).

Approximately 66 percent of the minimum global population of Kittlitz's murrelet is associated with glacially affected marine waters in the breeding season. Within these areas, Kittlitz's murrelets prefer highly stratified, cool, turbid marine waters near tidewater glaciers and glacial outflows, especially in the vicinity of submerged marine sills where localized upwelling occurs (Day and Nigro 2000, pp. 5, 8; Kissling
et al.
2007, pp. 2171-2172; Allyn
et al.
2012, p. 244; Arimitsu
et al.
2012, p. 18). The reason that Kittlitz's murrelets use these areas is not clear, but several hypotheses have been proposed. For example, marine waters with these characteristics may provide increased abundance of high-energy forage fish, such as sand lance or capelin (Robards
et al.
2003, p. 71; Arimitsu
et al.
2008, p. 137; Arimitsu
et al.
2011, pp. 15, 17-18; Renner
et al.
2012, pp. 2037-2038), or promote greater foraging efficiency for Kittlitz's murrelets (Day
et al.
2003, pp. 695-696; Arimitsu
et al.
2011, p. 14; Allyn
et al.
2012, pp. 244-245). Nutrient-rich glacial meltwater (Crusius
et al.
2011, p. 1) forms a turbid, stratified surface layer that limits light penetration, reducing phytoplankton

growth at depth (Hood
et al.
2009, p. 1046; Piwosz
et al.
2009, pp. 552-554, 556) and possibly affecting vertical diel (24 hour) migration of zooplankton and fish (Abookire
et al.
2002, p. 378; Frank and Widder 2002, p. 1189). Owing to their proportionately larger-diameter eye compared to the marbled murrelet (Day
et al.
2003, p. 695), the Kittlitz's murrelet may specialize at foraging in these low light conditions, taking advantage of underutilized ocean space and prey. In the northern Gulf of Alaska, freshwater streams and rivers fed by glaciers and snow melt drain into the coastal ocean and create large plumes of highly turbid water (Crusius
et al.
2011, pp. 1-2), where both zooplankton and juvenile fish abundance is greater compared to outside the plumes (McFadden
et al.
2012, p. 1). Juvenile fish may occupy these areas to take advantage of concentrated zooplankton populations or to evade predation (McFadden
et al.
2012, p. 1). Several studies have also suggested that the physical features and landforms (e.g., underwater sills and moraines) within glacial fjords interact with tides to concentrate prey of the Kittlitz's murrelet (Kissling
et al.
2007, p. 2171; Allyn
et al.
2012, pp. 244-245; Arimitsu
et al.
2012, pp. 10-15). Yet no studies have reported greater foraging success, or subsequent productivity or survival, in glacially affected waters compared to those without glacial influence, or in fjord versus non-fjord habitats (e.g., outer coast of the Gulf of Alaska).

Any foraging advantages in glacially affected waters should be readily apparent in the breeding season when Kittlitz's murrelets concentrate in these areas and deliver whole fish singly to chicks at nests. However, nests have been found throughout this species' range, including many areas without tidewater glaciers or glacially influenced marine waters (e.g., Kodiak and Aleutian islands, northern Alaska, and Russia), and, although highly variable, chick meal delivery rates at nests monitored at glacial sites (Naslund
et al.
1994, p. 46; Kissling, unpublished data) are not substantially different from those at non-glacial sites (Lawonn 2012, pp. 27-28, 55; Kaler, unpublished data), with one exception. Delivery rates for Agattu Island are much higher than those for all other sites, but the lack of glacial influence in the marine system alone cannot explain the unusually high rate of 10.2 fish per day, especially when compared to the moderate rate of 6.3 fish per day at nearby Adak Island (Kaler, unpublished data). Agattu Island is the only study site where rockfish and Pacific cod, low-energy-density fishes (Anthony
et al.
2000, p. 75), have been delivered as chick meals at monitored nests (Kaler, unpublished data), likely explaining the higher delivery rates and lower fledging mass of chicks. However, there is no information to suggest that the absence of high-quality fishes in the chick diet of Kittlitz's murrelets on Agattu Island is associated with the absence of glaciers in this region. On nearby Buldir Island in the western Aleutians, chick diets of tufted puffin (
Fratercula cirrhata
) and horned puffin (
F. corniculata
) between 1988 and 2012 were consistently composed of low-quality fish (i.e., hexagrammids) with intermittent years of relatively high percentages of high-quality Pacific sand lance (Warzybok
et al.
2013, pp. 162, 180). Therefore, although poor quality forage fish may be affecting nesting success of Kittlitz's murrelets on Agattu Island, it appears to be related to natural and regional fluctuations in forage fish abundance that cannot be attributed to the lack of glacial influence. Similarly, on non-glaciated Kodiak Island, the chick meal delivery rate (4.6 fish per day) is comparable to that estimated at glaciated sites (3-5 fish per day) (Naslund
et al.
1994, p. 46; Kissling, unpublished data). It is possible, but extremely unlikely, that Kittlitz's murrelets nesting on Kodiak Island make the lengthy round-trip flight to forage in the glacially-affected waters of Kenai Fjords (488 mi round-trip [784 km]), Kachemak Bay (374 mi [602 km]), Lower Cook Inlet (250 mi [402 km]), or perhaps to the far less-glaciated waters of the Alaska Peninsula (31 mi [50 km]) and then return to their nests with fish. For all of these reasons, we cannot determine whether glacially affected waters are a required or advantageous (in terms of fitness) element of breeding habitat for the Kittlitz's murrelet.

In addition to chick diet, trophic level and stomach contents of adult Kittlitz's murrelets sampled in the breeding season did not differ between glaciated and non-glaciated areas (Day
et al.
1999, p. 9). In glacial fjords of southeastern Alaska, adult Kittlitz's murrelets captured in the early breeding season (May) were heavier compared to those captured in the late breeding season (late July-August) (Kissling, unpublished data). Reduced body mass of Kittlitz's murrelets during the breeding season may be aimed at increasing flight efficiency and reducing energetic costs of transiting to and from nest sites (Hatch 2011, p. 82), but too few murrelets appear to attempt to breed annually (18 percent; see
Reproductive Performance,
above) (Kissling, unpublished data) to explain the overall change in body mass between early and late breeding periods. Furthermore, the rapid departure from breeding sites (Robards
et al.
2003, pp. 92, 100, 104; Kissling
et al.
2007, pp. 2167-2168; Madison
et al.
2012, p. 1) suggest that the foraging conditions and resources in glacially-affected waters are suitable and sufficient for breeding only for a short period. Otherwise, it is reasonable to assume that murrelets would remain in the

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A2013-24172. Public record. Not legal advice.