Draft Recovery Plan for Five Species from American Sāmoa

Home , Crow honeyeater, Emballonura, Eua (gastropod), Eua zebrina, Forest honeyeater, Neocyclotidae

U.S. Fish & Wildlife Service Draft Recovery Plan for Five Species From American Sāmoa

J. Malotaux R. Stirnemann Pe'ape'a Vai or Pacific Sheath-tailed BatBatBat Ma'oma'o Emballonura semicaudata semicaudata Gymnomyza samoensis

Tu'aimeo or Friendly Ground-Dove J. Malotaux Gallicolumba stairi

Eua zebrina R. RundellRundellR.

R. Rundell Ostodes strigatus Draft Recovery Plan for Five Species from American Sāmoa

Peʻapeʻa Vai or Pacific Sheath-tailed Bat, South Pacific Subspecies (Emballonura semicaudata semicaudata)

Maʻomaʻo or Mao (Gymnomyza samoensis)

Tuʻaimeo or Friendly Ground-Dove (Gallicolumba [=Alopecoenas] stairi) American Sāmoa Distinct Population Segment

Eua zebrina

Ostodes strigatus

June 2020 U.S. Fish and Wildlife Service Portland, Oregon

Approved: XXXXXXXXXXXXXXXXXXXX Regional Director U.S. Fish and Wildlife Service DISCLAIMER

Recovery plans delineate reasonable actions needed to recover and/or protect listed species. We, the U.S. Fish and Wildlife Service (Service), publish recovery plans, sometimes preparing them with the assistance of recovery teams, contractors, State agencies, and others. Objectives of the recovery plan are accomplished, and funds made available, subject to budgetary and other constraints affecting the parties involved, as well as the need to address other priorities with the same funds.

Recovery plans do not necessarily represent the views or the official positions or approval of any individuals or agencies involved in the plan formulation, other than our own. They represent our official position only after signed by the Director or Regional Director. Draft recovery plans are reviewed by the public and may be subject to additional peer review before the Service adopts them as final. Recovery objectives may be attained and funds expended contingent upon appropriations, priorities, and other budgetary constraints. Recovery plans are guidance and planning documents only; identification of an action to be implemented by any public or private party does not create a legal obligation beyond existing legal requirements. Nothing in this plan should be construed as a commitment or requirement that any Federal agency obligate or pay funds in any one fiscal year in excess of appropriations made by Congress for that fiscal year in contravention of the Anti-Deficiency Act, 31 U.S.C. 1341, or any other law or regulation. Approved recovery plans are subject to modification as dictated by new findings, changes in species status, and completion of recovery actions.

Literature citation of this document should read as follows: U.S. Fish and Wildlife Service. 2020. Draft recovery plan for five species from American Sāmoa. Portland, Oregon. xiii + 50 pages.

An electronic copy of this recovery plan is also available at: https://www.fws.gov/pacific/ecoservices/endangered/recovery/plans.html and

http://www.fws.gov/endangered/species/recovery-plans.html

i ACKNOWLEDGMENTS

We, the U.S. Fish and Wildlife Service, gratefully acknowledge the contributions of the following individuals and organizations in developing the Draft Recovery Plan for Five Species from American Sāmoa. The primary authors of this draft recovery plan are Joy Browning and Ryan Peʻa (Pacific Islands Fish and Wildlife Office [PIFWO], Honolulu, Hawaiʻi), and Grant Canterbury (Regional Office, Portland, Oregon).

We particularly thank the species experts who provided their input during development of the draft recovery plan:

Scott Burch, National Park of American Sāmoa Robert Cowie, Pacific Biosciences Research Center, University of Hawaiʻi at Mānoa Adam Miles, American Sāmoa Department of Marine and Wildlife Resources Alivereti Naikatani, University of the South Pacific David Sischo, Snail Extinction Prevention Program, Hawaiʻi Division of Forestry and Wildlife Rebecca Stirnemann, Australian National University Nunia Thomas-Moko, NatureFiji David Waldien, Kutztown University of Pennsylvania

ii RECOVERY PLANNING PROCESS

The Service’s Recovery Planning and Implementation (RPI) approach to recovery planning (see https://www.fws.gov/endangered/esa-library/pdf/RPI-Feb2017.pdf) is intended to reduce the time needed to develop and implement recovery plans, increase recovery plan relevancy over a longer timeframe, and add flexibility to recovery plans so they can be adjusted for new information or circumstances. Under the RPI process, a recovery plan includes the statutorily-required elements under section 4(f) of the Endangered Species Act (Act) (objective and measurable recovery criteria, site-specific management actions, and estimates of time and costs), along with a concise introduction and our strategy for how we plan to achieve species recovery. An RPI recovery plan is supported by two supplementary documents: a Species Status Assessment or Species Report, which describes the best available scientific information related to the biological needs of the species and assessment of threats; and the Recovery Implementation Strategy, which details the particular near-term activities needed to implement the recovery actions identified in the recovery plan. Under this approach, new information on species biology or details of recovery implementation may be incorporated by updating these supplementary documents without the need to revise the recovery plan, unless changes to statutorily-required elements are necessary.

Thus, this draft recovery plan document is one piece of a three-part framework:

1. The Recovery Plan contains a concise overview of the recovery strategy for the species (indicating how its recovered state will achieve redundancy, resiliency, and representation), as well as the statutorily required elements of recovery criteria, recovery actions, and estimates of the time and costs to achieve the plan’s goals.

2. The Species Status Assessment (SSA) or Species Report informs the Recovery Plan; it describes the biology and life history needs of the species, includes analysis of each subspecies’ historical and current conditions, and includes discussion of threats and conservation needs of each subspecies, as well as detailed literature references. The SSA or Species Report’s format is structured around the conservation biology principles of resiliency, redundancy, and representation (Shaffer and Stein 2000, pp. 307-310; Wolf et al. 2015, entire). Biology and threat status for the species addressed in this draft recovery plan, including the geography and environmental context of their range in the Samoan archipelago (and more broadly for the bat), are described in five Species Report documents (USFWS 2020a, 2020b, 2020c, 2020d, and 2020e).

3. The Recovery Implementation Strategy (RIS) is the vehicle for implementing the Recovery Plan. The RIS is a short-term, flexible operational document focused on how, when, and by whom the recovery actions from the Recovery Plan will be implemented. This approach allows us to incorporate new information and adapt to changing circumstances with greater flexibility and efficiency. The RIS will be developed and maintained in cooperation with our conservation partners, and will focus on the period of time and scope of activities that work best for our partners to achieve recovery goals. For this draft recovery plan, we are in the process of coordinating with conservation partners in American Sāmoa (particularly the National Park of American Sāmoa [NPSA] and the American Sāmoa Department of Marine and Wildlife Resources [DMWR]) and internationally to identify the highest priority actions for recovery of these species and finalize the draft RIS.

iii EXECUTIVE SUMMARY

Species Status This draft recovery plan addresses one mammal (peʻapeʻa vai, the South Pacific subspecies of Pacific sheath-tailed bat [Emballonura semicaudata semicaudata]); two birds (maʻomaʻo or mao [Gymnomyza samoensis] and the American Sāmoa distinct population segment (DPS) of tuʻaimeo or friendly ground-dove [Gallicolumba stairi]); and two snails (Eua zebrina and Ostodes strigatus). These five species were proposed for listing as endangered on October 13, 2015 (USFWS 2015) and were listed as endangered on September 22, 2016 (USFWS 2016). No critical habitat is designated for these species. All of these species are currently or historically known from American Sāmoa; the peʻapeʻa vai and maʻomaʻo also occur internationally. Recovery Vision Recovery of peʻapeʻa vai will be signified by restoration of multiple redundant populations distributed widely throughout its international range, comprising one or more viable and interconnected metapopulations allowing natural processes of dispersal and recolonization. Populations should be resilient and self-sustaining, with threats effectively managed through protection of roosts and foraging habitat, and conserve existing genetic diversity. Recovery of maʻomaʻo will be signified by the presence of protected, resilient local populations conserving existing genetic diversity (with sufficient habitat and with predation pressure adequately managed to be self-sustaining), redundantly on multiple islands in Sāmoa and American Sāmoa. Recovery of tuʻaimeo will be signified by the presence of protected, resilient local populations conserving existing genetic diversity (with sufficient habitat and with predation pressure adequately managed to be self-sustaining), redundantly on multiple islands in American Sāmoa. Recovery of Eua zebrina and Ostodes strigatus will be signified by the presence of multiple redundant local populations in American Sāmoa that are resilient, self-sustaining, and conserve existing genetic diversity. Populations must not be significantly threatened by nonnative predators or other threats (either because they exist on islands where predators have not been introduced, or because the impact of predation does not threaten their continued existence).

Recovery Strategy Major recovery strategy elements include:

• For all species: protection of native forest habitat; biosecurity measures to prevent transportation of nonnative predators and invasive plants among islands in American Sāmoa and introduction of new invasive species from outside the territory; and modification of American Sāmoa territorial law to allow eligibility for section 6 grant funds under the Act.

• For peʻapeʻa vai: protection of roost sites and nearby forest from disturbance and habitat modification (particularly in the known remaining populations in Fiji); rangewide surveys

to locate other extant populations; research on potential pesticide and disease impacts; and evaluating the feasibility of reestablishing populations through translocation or other approaches.

• For maʻomaʻo: predator control programs for rats and cats in Sāmoa (as well as any translocation localities in American Sāmoa); protection of diverse mature forest; and if feasible, translocation to reestablish populations in American Sāmoa.

• For tuʻaimeo: predator control programs for rats and cats on Ofu, Olosega, and any translocation localities; maintaining protected forest habitat; and if feasible, translocation to establish populations on other islands.

• For Eua zebrina and Ostodes strigatus: comprehensive survey of American Sāmoa islands and islets for native snail and predator populations; translocation to predator-free islands and/or establishment of captive propagation programs; evaluating demography and predator-prey dynamics in existing populations; assessment of genetic and morphological diversity within and among islands; developing new predator control methods; and maintaining protected forest habitat. We note that new or emerging threats may arise over time. If any such issues are so significant as to result in a need to change the recovery strategy or the statutorily required elements of the recovery plan under section 4(f)(1) of the Act, we will amend or revise the plan as appropriate.

Recovery Criteria Recovery criteria represent our best assessment, at the time the recovery plan is completed, of the conditions that would likely result in a determination that listing under the Act as threatened or endangered is no longer required. However, revisions to the Lists, including delisting or downlisting a species, must reflect determinations made in accordance with sections 4(a)(1) and 4(b) of the Act. Section 4(a)(1) requires that the Secretary determine whether a species is an endangered species or threatened species because of threats to the species, based on an analysis of the five listing factors in section 4(a)(1). Section 4(b) require that the determination be made “solely on the basis of the best scientific and commercial data available.” Thus, while recovery plans provide important guidance to the Service, States, and other partners on methods of minimizing threats to listed species and measurable criteria against which to measure progress towards recovery, they are guidance and not regulatory documents. Thus, a decision to delist or downlist a species is informed by the recovery criteria but is ultimately based on an analysis of threats using the best scientific and commercial data then available.

Peʻapeʻa vai: Downlisting Criteria:

surveys of abundance or an index of abundance derived from quantitative surveys or demographic monitoring. The average population over this time period is estimated to be at least 5,000 individuals in each management unit.

(2) Roosts that contribute to Downlisting Criterion 1 are surrounded by suitable forest foraging habitat and are being effectively protected from disturbance, with long-term management commitments in place. In Fiji, protection from disturbance includes at least those roost sites currently or historically occupied by large populations, including those on Taveuni, Vatu Vara, and Yaqeta islands.

(3) Threats to the species (including impacts of predation, habitat alteration, and pesticides) are being effectively managed such that mortality is reduced and populations meet targets in Downlisting Criterion 1. Delisting Criteria:

(1) Peʻapeʻa vai populations are established and stable in Fiji and at least 2 other management units over a period of an additional 10 years, with active, consistently occupied roosts represented on multiple islands in each management unit. Peʻapeʻa vai population data in each management unit show a stable or positive growth curve, as determined through quantitative surveys of abundance or an index of abundance derived from quantitative surveys or demographic monitoring. The average population throughout that time period is estimated to be at least 10,000 individuals in each management unit.

(2) Roosts that contribute to Delisting Criterion 1 are surrounded by suitable forest foraging habitat and are being effectively protected from disturbance, with long-term management commitments in place. Roost site protection includes areas described under Downlisting Criterion 2 as well as expansion to additional management units.

Maʻomaʻo: Downlisting Criteria:

(1) Over a period of at least 20 years, maʻomaʻo population data show a stable or increasing trend (i.e., finite rate of population increase, or lambda, greater than or equal to 1) that is statistically significant, as determined through quantitative surveys of abundance, or an index of abundance derived from quantitative surveys or demographic monitoring; and the average population throughout that time period is estimated to be at least 500 individuals on both Savaiʻi and ʻUpolu.

(2) Mature forest habitat in Sāmoa is protected with long-term management commitments and habitat restoration has been completed to the extent that the amount of habitat

iii

suitable for maʻomaʻo is sufficient to sustain population targets identified in Downlisting Criterion 1.

(3) Predator control programs in maʻomaʻo habitat on Savaiʻi and ʻUpolu are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Downlisting Criterion 1 are met. Delisting Criteria: (1) Over a period of an additional 20 years, maʻomaʻo population data show a stable or increasing trend (i.e., finite rate of population increase, or lambda, greater than or equal to 1) that is statistically significant, as determined through quantitative surveys of abundance, or an index of abundance derived from quantitative surveys or demographic monitoring. The average population throughout that time period is estimated to be at least 1,000 individuals on both Savaiʻi and ʻUpolu; or the average population is estimated to be at least 750 individuals on both Savaiʻi and ʻUpolu, and one or more self-sustaining populations totaling at least 250 individuals are established within American Sāmoa. (2) Mature forest habitat in Sāmoa and at any translocation localities in American Sāmoa is protected and restoration has been completed to the extent that the amount of habitat suitable for maʻomaʻo is sufficient to sustain population targets identified in Delisting Criterion 1. (3) Predator control programs in maʻomaʻo habitat on Savaiʻi, ʻUpolu, and any translocation localities in American Sāmoa are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Delisting Criterion 1 are met.

Tuʻaimeo: Downlisting Criteria: (1) Over a period of at least 10 years, tuʻaimeo population data show a stable or increasing trend (i.e., finite rate of population increase, or lambda, greater than or equal to 1) that is statistically significant, as determined through quantitative surveys of abundance, or an index of abundance derived from quantitative surveys or demographic monitoring; and the average population throughout that time period is estimated to be at least 250 individuals on at least 1 island (Ofu or Olosega). (2) Forest habitat on Ofu and Olosega is protected and managed such that the amount of habitat suitable for tuʻaimeo is sufficient to sustain population targets identified in Downlisting Criterion 1. (3) Predator control programs on Ofu and Olosega are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Downlisting Criterion 1 are met.

Delisting Criteria:

(1) Over a period of an additional 10 years, tuʻaimeo population data show a stable or increasing trend (i.e., finite rate of population increase, or lambda, greater than or equal to 1) that is statistically significant, as determined through quantitative surveys of abundance, or an index of abundance derived from quantitative surveys or demographic monitoring. The average population throughout that time period is estimated to be at least 250 individuals on at least 2 islands within American Sāmoa (including Ofu, Olosega, or potentially Taʻu or Tutuila). (2) Forest habitat on Ofu and Olosega and any translocation localities is protected and managed such that the amount of habitat suitable for tuʻaimeo is sufficient to sustain population targets identified in Delisting Criterion 1.

(3) Predator control programs on Ofu, Olosega, and any translocation localities are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Delisting Criterion 1 are met.

Eua zebrina and Ostodes strigatus: Downlisting Criteria:

(1) At least six stable populations (possibly actively managed) exist on one or more islands within American Sāmoa. To be considered stable, a population must number at least 500 individuals distributed across all size classes combined, and must have a population growth curve or index trend that is stable or positive for at least 4 of 5 sequential years. If multiple management subunits have been identified for the species based on morphological or genetic characters, each subunit must comprise one or more of these stable populations.

(2) Each population in Downlisting Criterion 1 occurs on suitable habitat that is managed to protect native forest vegetation.

(3) Biosecurity measures are in place to prevent predator introduction to new islands, and evaluation of predation risk for each population in Downlisting Criterion 1 indicates either that nonnative predators are absent or that predation is unlikely to have significant short-term impacts on the population. Delisting Criteria:

(1) At least 12 free-living populations exist on two or more islands within American Sāmoa. Each population must number at least 500 individuals, distributed across all size classes combined, and must have a population growth curve or index trend that is stable or positive for at least 7 of 10 sequential years. If multiple management subunits have been identified for the species based on morphological or genetic characters, each subunit must comprise two or more of these populations. (2) Each population in Delisting Criterion 1 occurs in suitable habitat that is managed to protect native forest vegetation and is capable of supporting expansion of the occupied

range, positive population growth, and establishment of new populations through natural dispersal. (3) Biosecurity measures are in place to prevent predator introduction to new islands, and predation does not threaten long-term viability of the populations in Delisting Criterion 1 because a) nonnative snail predators do not exist on the island or islet where the population occurs; b) effective predator control with long-term management commitment has successfully reduced predation pressure such that population viability is maintained; or c) quantitative data on demography and predator/prey dynamics in the population indicate that the population will maintain long-term viability without active predator control. At least 6 of the populations in Delisting Criterion 1 must exist on islands or islets where surveys confirm Euglandina (rosy wolf-snail) and Platydemus manokwari (New Guinea flatworm) are absent.

Recovery Actions and Estimated Costs

Recovery Actions Estimated Cost 1.0 Survey and protect management units. $21M

2.0 Control species-specific threats. $106M+ 3.0 Expand existing wild populations and establish additional populations. $1.7M 4.0 Develop regulations and policy to support species recovery. $180M

5.0 Conduct additional research. $1M Estimated cost through date of recovery (25-60 years by species, as discussed below): $310M+

Cost estimates are preliminary. Project-level details of recovery action implementation are to be developed cooperatively with partners in a Recovery Implementation Strategy (RIS) for this recovery plan. Implementation is subject to availability of funds and is at the discretion of partners and sovereign governments.

Date of Recovery: Recovery criteria could be met by the year 2060 for peʻapeʻa vai, 2080 for maʻomaʻo, 2050 for tuʻaimeo, and 2045 for Eua zebrina, based on the conservative assumption that all recovery plan actions are fully funded and implemented as currently outlined, including gaining full cooperation of all partners needed to achieve recovery. Recovery of Ostodes strigatus cannot be reasonably projected unless a population is rediscovered, but in any scenario recovery is unlikely to be achieved before the year 2060.

ACRONYMS AND ABBREVIATIONS DMWR: American Sāmoa Department of Marine and Wildlife Resources

DPS: Distinct Population Segment

ESA or Act: Endangered Species Act

Fiji: Republic of Fiji

MNRE: Sāmoa Ministry of Natural Resources and Environment

NMFS: National Marine Fisheries Service

NPSA: National Park of American Sāmoa

RIS: Recovery Implementation Strategy

Sāmoa: Independent State of Sāmoa

Tonga: Kingdom of Tonga

USFWS: U.S. Fish and Wildlife Service

Vanuatu: Republic of Vanuatu

vii

TABLE OF CONTENTS

I. INTRODUCTION ...... 1 A. Background ...... 1 1. Basic Species Information ...... 1 2. Threats ...... 7 II. RECOVERY ...... 10 A. Recovery Vision and Strategy ...... 10 1. Recovery Vision ...... 10 2. Recovery Strategy ...... 10 B. Recovery Criteria ...... 21 1. Peʻapeʻa Vai ...... 21 2. Maʻomaʻo ...... 24 3. Tuʻaimeo ...... 26 4. Eua zebrina and Ostodes strigatus ...... 29 III. RECOVERY ACTIONS ...... 32 IV. TIME AND COST ESTIMATES ...... 42 V. REFERENCES ...... 45 A. Literature Cited ...... 45 B. IN LITT. References ...... 50

LIST OF FIGURES Figure 1. Sāmoa, Fiji, and Tongan archipelagos ...... 2

Figure 2. Samoan archipelago ...... 2

Figure 3. Current and historical roost locations of the Pacific sheath-tailed bat in Fiji ...... 3

Figure 4. Historical roost locations of the Pacific sheath-tailed bat in Tonga ...... 4

LIST OF TABLES Table 1. Current and historical status and distribution of the five endangered species ...... 6 Table 2. Summary of habitats used by, and the threats affecting, the five listed species found in American Sāmoa ...... 9

Table 3. Management units by national boundaries ...... 11

viii

Table 4. Management subunits by islands within American Sāmoa ...... 11

Table 5. Crosswalk relating threats, recovery criteria, and recovery actions ...... 39

Table 6. Priority and estimated costs of recovery actions ...... 43

I. INTRODUCTION

The Endangered Species Act of 1973, as amended (16 U.S.C. 1531 et seq.) (Act) protects species of wildlife and plants that are listed as endangered or threatened. Recovery is defined as “the process by which listed species and their ecosystems are restored and their future is safeguarded to the point that protections under the Act are no longer needed”, according to the 2018 updated National Marine Fisheries Service (NMFS) and U.S. Fish and Wildlife Service (Service) Interim Recovery Planning Guidelines, Version 1.4 (NMFS and USFWS 2018).

Recovery plans are guidance documents developed to provide recommendations to reduce or alleviate threats to the species and ensure self-sustaining populations in the wild. The Act (section 4(f)(1)) stipulates that recovery plans shall, to the maximum extent practicable, incorporate: “(i) a description of such site-specific management actions as may be necessary to achieve the plan’s goal for the conservation and survival of the species; (ii) objective, measurable criteria which, when met, would result in a determination, in accordance with the provisions of this section, that the species be removed from the list [Federal Lists of Endangered and Threatened Wildlife and Plants]; and (iii) estimates of the time required and the cost to carry out those measures needed to achieve the plan’s goal and to achieve intermediate steps toward that goal.”

This draft recovery plan addresses one mammal (peʻapeʻa vai, the South Pacific subspecies of Pacific sheath-tailed bat [Emballonura semicaudata semicaudata]); two birds (maʻomaʻo or mao [Gymnomyza samoensis] and the American Sāmoa distinct population segment (DPS) of tuʻaimeo or friendly ground-dove [Gallicolumba stairi]); and two snails (Eua zebrina and Ostodes strigatus). These five species were proposed for listing as endangered on October 13, 2015 (USFWS 2015) and were listed as endangered on September 22, 2016 (USFWS 2016). No critical habitat is designated for these species. All of these species are currently or historically known from American Sāmoa (Figures 1 and 2); the peʻapeʻa vai and maʻomaʻo also occur internationally. A recovery outline describing a preliminary course of action for meeting recovery needs of these species was completed on December 9, 2019 (USFWS 2019a). The first 5-year status reviews for these species were initiated in 2019 (USFWS 2019b) and are currently in development.

A. BACKGROUND 1. Basic Species Information

For a more detailed literature review and full description of the biology, habitat, and distribution of the five species addressed in this draft recovery plan, please refer to the proposed and final listing rules (USFWS 2015, 2016) and the five Species Reports (USFWS 2020a, 2020b, 2020c, 2020d, and 2020e) (see Species Description, Historical Range, and Current Range and Distribution sections). Current status and distribution are summarized in Table 1 below.

Figure 1. Sāmoa, Fiji, and Tonga archipelagos (USFWS 2020a).

Figure 2. Samoan archipelago (USFWS 2020a, b, c, d, e).

The Pacific sheath-tailed bat (Emballonura semicaudata) is a nocturnal insectivorous bat that roosts colonially in caves and forages in nearby forested habitat. It is widely distributed on Pacific islands in Polynesia, Micronesia, and Melanesia (Hutson et al. 2001, pp. 138-139). Four subspecies have been described. This draft recovery plan addresses the South Pacific subspecies, with the Sāmoan name peʻapeʻa vai and the Fijian name bekabeka (Emballonura semicaudata semicaudata), which historically occurred in the Samoan archipelago, Tonga, Fiji, and Vanuatu (Hutson et al. 2001, pp. 138-139; USFWS 2020a). The primary remaining populations are in Fiji (Figure 3); it has not been observed in American Sāmoa since 1998 and is very rare or extirpated in Tonga (Figure 3), Sāmoa, and Vanuatu (Helgen and Flannery 2002, pp. 209-211; Hutson et al. 2001, pp. 138-139; Koopman and Steadman 1995, p. 7; Palmeirim et al. 2007, pp. 512-517; Scanlon et al. 2014, pp. 453-456; Tarburton 2002, pp. 106-107).

Figure 3. Current and historical roost locations of the Pacific sheath-tailed bat in Fiji (USFWS 2020a).

3 Figure 4. Historical roost locations of the Pacific sheath-tailed bat in Tonga (USFWS 2020a).

The maʻomaʻo, or mao (Gymnomyza samoensis), is a large honeyeater (a bird in the family Meliphagidae) endemic to the Samoan archipelago (Ministry of Natural Resources and Environment [MNRE] 2006, p. 1; USFWS 2020b, pp. 18-21). It inhabits mature forests and forages primarily in the canopy and subcanopy for nectar, invertebrates, and fruit (Butler and Stirnemann 2013, p. 30). Populations are currently known only from the islands of ʻUpolu and Savaiʻi in Sāmoa (MNRE 2006, pp. 2-4). A few records exist from the island of Tutuila in American Sāmoa, but it has not been reported since 1977 (Amerson et al. 1982, p. 72; Pratt et al. 1987, p. 280). Because the species is loud and vocal, it is highly unlikely that an undetected population could persist in American Sāmoa and it is considered extirpated from the territory (Engbring and Ramsey 1989, p. 68).

The friendly ground-dove (Gallicolumba [=Alopecoenas] stairi) is a medium-sized dove that occurs in the Samoan archipelago, Fiji, Tonga, and the Wallis and Futuna Islands (Watling 2001, pp. 117-118). It inhabits forested and brushy habitats, foraging on the ground and in the understory for seeds, fruit, and invertebrates (Clunie 1999, pp. 42-44). This draft recovery plan addresses the American Sāmoa DPS of the species, with the Samoan name of tuʻaimeo, which

4 occurs only on the islands of Ofu and Olosega (USFWS 2016, p. 65490; USFWS 2020c, pp. 17- 22).

Eua zebrina is a tree snail in the family Partulidae (USFWS 2020d) that occurs in forested habitat, primarily on leaves but also on trunks and branches and in leaf litter (Cowie 1992, p. 175). It is known only from the islands of Tutuila and Ofu in American Sāmoa. The most recent systematic surveys for American Sāmoa snails were conducted in 1992 on Tutuila, Ofu, Olosega, and Ta‘u, reported by Miller (1993); in 1998 primarily within National Park Service American Sāmoa [NPSA] boundaries on Tutuila, Ofu, and Ta‘u, reported by Cowie and Cook (1999, 2001); and in 2001 on Olosega (reported by Cowie et al. 2003). These surveys documented Eua zebrina at multiple localities on Tutuila in 1992 and 1998, on Nuʻusetoga islet near Tutuila in 1992, and at one locality on Ofu in 1998. Subsequent information from site-specific surveys within the NPSA and anecdotal observations indicates that Eua zebrina continues to be regularly observed and remains extant on Tutuila, but we do not have any systematic current information about its distribution or abundance, or whether these have changed since 1998 (A. Miles in litt. 2015). No new information is available about the Ofu population.

Ostodes strigatus is a land snail in the family Neocyclotidae (in some classifications Poteriidae) that occurs on the ground in leaf litter within closed-canopy forests (Cowie 1998, p. 24, USFWS 2020e). It is known only from the western portion of the island of Tutuila in American Sāmoa (Girardi 1978, p. 224). The last survey to detect the species, in 1992, found abundant empty shells in its former range but detected only a few live individuals, restricted to a single site (Miller 1993, pp. 11, 24-27). There have been no known observations of the species since (A. Miles in litt. 2015), and it may be extinct.

5 Table 1. Current and historical status and distribution of the five endangered species across the four islands comprising the Territory of American Sāmoa, and elsewhere.

Current and Historical Distribution Species # of # of Tutuila Ofu Olosega Taʻu Other populations individuals MAMMALS Peʻapeʻa vai unknown H3 H3 H3 H3 Sāmoa (H) Bekabeka Fiji (✓) Pacific sheath-tailed bat Tonga (H) (Emballonura Vanuatu (H) semicaudata semicaudata) (USFWS 2020a) BIRDS Maʻomaʻo 7 <500 H4 Sāmoa (✓) Mao (2006) (Gymnomyza samoensis) (USFWS 2020b) Tuʻaimeo 2 <100 ✓ ✓ Friendly ground-dove1 (2015) (Gallicolumba stairi) (USFWS 2020c) SNAILS Eua zebrina2 2 1,102 ✓ ✓ (USFWS 2020d) (1998) Ostodes strigatus2 Unknown H5 (USFWS 2020e) ✓ = Currently known populations H = Known only historically 1 = Tuʻaimeo: American Sāmoa DPS (the species is also found on the other islands of the Samoan, Fijian, and Tongan archipelagos, and Wallis and Futuna Islands). 2 = Both Eua zebrina and Ostodes strigatus are endemic to American Sāmoa 3 = Peʻapeʻa vai : Colonies were reported on all major American Sāmoa islands by Amerson et al. (1982), but none are currently known to be extant. The last individuals seen were on Tutuila in 1998. It is rare or extirpated in Sāmoa and Tonga, and populations in Fiji have declined significantly in recent years. 4 = Maʻomaʻo: Last collected in 1920 on Tutuila. Unconfirmed sightings are occasionally reported on Tutuila. 5 = Ostodes strigatus: Last surveyed and observed in 1992 on Tutuila.

2. Threats

For a more detailed literature review and full description of threats to the five species addressed in this draft recovery plan, please refer to the proposed and final listing rules (USFWS 2015, 2016) and the five Species Reports (USFWS 2020a, 2020b, 2020c, 2020d, and 2020e) (see Factors Affecting Viability sections). Species-specific threats are summarized in Table 2 below.

Major threats to the peʻapeʻa vai include degradation of forest habitat from deforestation, agricultural/urban development, and storms; direct mortality from storms; disturbance within roosting caves by humans and goats; forest clearance surrounding cave entrances and other roost areas; cyclone and sea-level rise impacts to coastal roosting caves; potential predation by cats or rats; and potential pesticide impacts (Grant et al. 1994, p. 135; Hansen et al. 2013; Hutson et al. 2001, p. 101; Meyer et al. 2017b, pp. 22-24; Palmeirim et al. 2005, pp. 3, 32-36, 46; 2007, p. 515; Tarburton 2002, pp. 106-107; Wiles et al. 2011, pp. 306-307; Wiles and Worthington 2002, p. 13; USFWS 2020a). Habitat impacts, coupled with the decline or extirpation of previously existing populations on many islands across its original range, likely disrupt natural processes of dispersal and recolonization (Palmeirim et al. 2005, p. 36).

Major threats to the maʻomaʻo and tuʻaimeo include loss or degradation of forest habitat resulting from agricultural/urban development, cyclone impacts, invasive plants, or pig activity; predation on nests, fledglings, and/or adult birds by rats (Rattus rattus, R. norvegicus, or R. exulans) and cats; and direct mortality from storms (ASCC 2010, pp. 13-15, 22; Atherton and Jefferies 2012, pp. 101-106; Butler and Stirnemann 2013, pp. 22, 28-31; Craig 2009, pp. 94-98; DMWR 2006, pp. 22-23; Kayano et al. 2019, p. 19; MNRE 2006, pp. 2, 4-10; Space and Flynn 2002, entire; Steadman and Freifeld 1998, p. 617; USFWS 2020b, pp. 22-30; USFWS 2020c, pp. 22-29; Watling 2001, p. 118.).

The primary threat to Eua zebrina and Ostodes strigatus is predation by nonnative predators, including the New Guinea flatworm (Platydemus manokwari), the rosy wolf-snail (Euglandina rosea complex) and other predatory snails (e.g., Gonaxis kibweziensis, Streptostele musaecola, and Gulella bicolor), and rats (Cowie and Cook 1999, pp. 36-37, 46-47; Cowie 2001, p. 219; Craig 2009, p. 84; Eldredge 1988, pp. 122-125; Miller 1993, pp. 10, 13-16, 23-28; USFWS 2020d, pp. 16-24; USFWS 2020e, pp. 19-26). Because of uncertainty about taxonomy of invasive Euglandina populations on Pacific islands (Meyer et al. 2017a, pp. 1403-1404), we collectively refer to them as Euglandina in this document. Euglandina and P. manokwari have colonized islands around the Pacific Ocean, resulting from both intentional introduction (in attempts to control an agricultural pest, the giant African snail [Achatina fulica]) and accidental transport (Clarke et al. 1984, pp. 97-104; Eldredge 1988, pp. 122-125; Hadfield et al. 1993, pp. 616-620; Ohbayashi et al. 2007, pp. 483-484). Predation by P. manokwari and Euglandina has been confirmed as a primary cause for dramatic declines and extinctions of native snails throughout this region (Cowie 1992, pp. 170-175). Snails in the family Partulidae (and Achatinellidae in the Hawaiian Islands) have proven particularly vulnerable to predation, likely because they are relatively slow-growing and slow-reproducing (Cowie 2001, p. 218).

Platydemus manokwari had been reported from Tutuila and Taʻu as of 2004 (Craig 2009, p. 84), but whether and where it is currently established is unknown. Euglandina has been established

on Tutuila since the 1980s (Eldredge 1988, pp. 122-125); it was reported from Taʻu in 1992 but was not observed in the 1998 survey (Miller 1993, p. 10; Cowie and Cook 1999, pp. 87-91). On Nuʻusetoga islet near Tutuila no Euglandina were observed (Miller 1993, p. 13). Euglandina is not known to occur on Ofu or Olosega (Cowie and Cook 1999, pp. 93-98; Cowie et al. 2003, p. 39).

It should be noted that in Hawaiʻi, introduced Euglandina populations previously identified as E. rosea have recently been found to comprise two distinguishable clades, potentially separate species, which may differ in the severity of their impacts on native snails. Only one of these Euglandina clades is known to have been introduced to American Sāmoa (Meyer et al. 2017a, p. 1402).

Impacts to forest habitat resulting from agricultural/urban development, invasive plants, or pig activity could also adversely affect microclimate or foraging substrates for Eua zebrina and Ostodes strigatus (Cowie 2001, p. 219).

Note that a similar suite of threats is likely to apply to various other snail species in American Sāmoa. Snail species of particular conservation concern in American Sāmoa include the three other non-listed partulid snails (Samoana abbreviata and S. conica on Tutuila and S. thurstoni on Ofu and Olosega) and Ostodes adjunctus (native to eastern Tutuila) (Cowie 1998, pp. 25, 53-55).

Table 2. Summary of habitats used by, and the threats affecting, the five listed species found in American Sāmoa.

Listing Factor A Factor B Factor C Factor D Factor E

) &

Species

Habitat (showing various names) rats cats Pigs Goats plants existing riculture F looding nonnative nonnative Nonnative regulatory mechanisms ag Predationby Predationby Predationby invertebrates Deforestation development Inadequacy of of Inadequacy Otherspecies Sea LevelRise ( specificthreats Overutilization

MAMMALS Peʻapeʻa vai HD ( Cave, ✓✓) Bekabeka H ( Forest ✓✓) Pacific sheath-tailed bat LN ( (Lowland, ✓ ✓ ✓✓ ✓✓ ✓ ✓ ✓ ✓✓) (Emballonura semicaudata Montane) MP ( semicaudata) ✓✓) P (✓) BIRDS Forest Maʻomaʻo (Lowland, H ( Mao ✓) Montane, ✓✓ ✓ ✓ ✓✓ ✓✓ ✓ (Gymnomyza samoensis) LN (✓✓) Cloud) Tuʻaimeo Forest H ( Friendly ground-dove1 (Lowland, ✓) ✓ ✓ ✓ ✓✓ ✓✓ ✓ (Gallicolumba stairi) Montane) LN (✓) SNAILS Forest EU ( H ( Eua zebrina (Lowland, ✓✓) ✓) ✓ ✓ ✓ ✓ ✓ ✓ ✓ Montane) FW (✓✓) LN (✓) Forest EU ( H ( Ostodes strigatus (Lowland, ✓✓) ✓) ✓ ✓ ✓ ✓ ✓ ✓ Montane) FW (✓✓) LN (✓✓) Factor A = The present or threatened destruction, modification, or curtailment of the species’ habitat or range. Factor B = Overutilization for commercial, recreational, scientific, or educational purposes. Factor C = Disease or predation. Factor D = Inadequacy of existing regulatory mechanisms. Factor E = Other natural or manmade factors affecting the species’ continued existence. Double check marks indicate threats that appear most significant to the species, based on currently available information. EU = Euglandina rosea, FW= New Guinea flatworm, H = hurricanes, HD = human disturbance, LN = low number of individuals, MP = breakdown of metapopulation dynamics via fragmentation, loss of subpopulations, etc., SL = sea-level rise, P = pesticides

II. RECOVERY

A. RECOVERY VISION AND STRATEGY

A recovery vision is an explicit expression of recovery in terms of resiliency (the ability of a species to recover from periodic disturbance), redundancy (the number of populations of a species distributed across the landscape), and representation (the range of variation found within a species). It builds upon the description of viability for the species and defines what recovery looks like for the species. The recovery strategy provides a recommended approach for achieving the recovery vision, and ultimately, the down- and delisting criteria.

1. Recovery Vision

Recovery of peʻapeʻa vai will be signified by restoration of multiple redundant populations distributed widely throughout its international range, comprising one or more viable and interconnected metapopulations allowing natural processes of dispersal and recolonization. Populations should be resilient, self-sustaining, and conserve existing genetic diversity, with threats effectively managed through protection of roosts and foraging habitat.

Recovery of maʻomaʻo will be signified by the presence of protected, resilient local populations conserving existing genetic diversity (with sufficient habitat and with predation pressure adequately managed to be self-sustaining), redundantly on multiple islands in Sāmoa and American Sāmoa.

Recovery of tuʻaimeo will be signified by the presence of protected, resilient local populations conserving existing genetic diversity (with sufficient habitat and with predation pressure adequately managed to be self-sustaining), redundantly on multiple islands in American Sāmoa.

Recovery of Eua zebrina and Ostodes strigatus will be signified by the presence of multiple redundant local populations in American Sāmoa that are resilient, self-sustaining, and conserve existing genetic diversity. Populations will not be significantly threatened by nonnative predators or other threats (either because they exist on islands where predators have not been introduced, or because sufficient management is in place that the impact of predation does not threaten their continued existence).

2. Recovery Strategy

Management Units For purposes of guiding action implementation, we identify management units based on national boundaries for species whose distribution extends beyond American Sāmoa (Table 3) and by islands where a species or DPS is endemic to American Sāmoa (Table 4). Units include various small islets in the vicinity of the major islands (e.g., Aunuʻu, Nuʻusetoga, and other islets near Tutuila; Nuʻutele, Nuʻulua, and other islets near ʻUpolu in Sāmoa).

10 Two management subunits for Eua zebrina are provisionally identified on the islands of Tutuila and Ofu; subunit boundaries should be further evaluated and geographically delineated based on assessment of morphological and genetic variation within and among Eua zebrina populations on Tutuila and Ofu, and used in planning of management and translocation activities.

Table 3. Management units by national boundaries. American Sāmoa Sāmoa Fiji Vanuatu Tonga Peʻapeʻa vai [extirpated] [extirpated] extant1 [extirpated] [possibly extirpated] Maʻomaʻo [extirpated] extant 1 = primarily in small Lau Group islands, and possibly on Taveuni or Rotuma

Table 4. Management subunits by islands within American Sāmoa. Tutuila Ofu Olosega Taʻu Tuʻaimeo (DPS) extant extant Eua zebrina extant extant Ostodes strigatus [possibly extirpated]

General Cross-Species Recovery Strategy Because all five of the species addressed in this draft recovery plan utilize forest habitat, protecting existing areas of mature forest provides benefits for all of the species in common. Forest ecosystem protection can be enhanced through incorporating sustainable forest management practices into management plans or conservation agreements with landowners in forest habitat; monitoring and control of invasive plant species; reducing pig impacts through lethal control or fencing; and minimizing expansion of commercial agricultural operations into existing forest habitat. The likely trends in localized microclimate (humidity and temperature) and forest vegetation composition should be better understood through local and regional stepdown modeling to project climate change impacts. Cyclones are expected to continue to intermittently cause catastrophic disruptions of forest habitat; while management is unlikely to prevent these impacts on any particular island, it is possible to mitigate the effects by maintaining forest habitats that include viable populations of the listed species on multiple islands in order to improve the redundancy of the species populations.

Because nonnative species (particularly predators and invasive plants) already negatively affect all of the species addressed in this recovery plan, biosecurity measures to prevent their transportation among islands in American Sāmoa are critical to avoid their establishment on previously unoccupied islands. Moreover, it is important to prevent additional new invasive species from being brought into American Sāmoa from outside the territory (in particular, brown tree snake [Boiga irregularis], which became established on Guam with catastrophic effects for the island’s birds). Such measures should include public outreach, coordination and inspection at ports of entry, and territorial biosecurity legislation.

The Territorial Endangered Species Act in American Sāmoa (American Sāmoa Code, Annotated Title 24, Chapter 7) authorized a commission governing the territorial listing of endangered species. Modification of territorial law to clarify agency authority and regulatory process for

11 protection of listed species, with establishment of a program to carry out this mandate, would allow the Territory to establish a cooperative agreement with the Service and to become eligible for funding of endangered species projects under section 6 of the Endangered Species Act. Access to such funds could support and enable a variety of the recovery actions identified below.

Recovery Strategy for Peʻapeʻa vai Our objective is to manage threats to enhance and reestablish self-sustaining, reproductively viable populations of peʻapeʻa vai in multiple island groups across its historical range. It is likely that multiple threats (including foraging habitat degradation, roosting disturbance, catastrophic storm events, and predation) have contributed to the decline and extirpation of local populations through much of the subspecies’ range, disrupting metapopulation dynamics of natural dispersal and recolonization.

Pacific sheath-tailed bats forage around their roost sites within mature well-structured forest with high and dense canopy (Esselstyn et al. 2004, pp. 306-307; Palmeirim et al. 2007, p. 515). Proximity of forest canopy is associated with occupation of roosts, as it can provide foraging habitat, function to moderate roost microclimatic conditions and reduce undergrowth, and visually shield the site and reduce disturbance (O’Shea and Valdez 2009, pp. 18, 44; Palmeirim et al. 2007, p. 515). In other species changes in cave microclimate have been associated with abandonment of bat roosts (Hutson et al. 2001, p. 101). Thus, protection of a matrix of mature forest ecosystems in the vicinity of roosting areas (e.g., through management plans, conservation agreements, invasive plant control) will be useful for maintaining viable foraging habitat and aid in protection or restoration of bat populations.

The low reproductive potential of sheath-tailed bats (one pup per female annually) leaves the species generally vulnerable to mortality events and reproductive failure and slow to recover from catastrophic events (Wiles et al. 2011, p. 303). Because cats or rats may opportunistically prey on roosting sheath-tailed bats (Wiles and Worthington 2002, p. 18; Wiles et al. 2011, p. 306; Palmeirim et al. 2007, p. 515), predator control measures could reduce this source of mortality. There may be opportunities to combine this management with predator control efforts being undertaken for other species.

Because peʻapeʻa vai are nocturnal and roost primarily within caves, they are vulnerable to harassment from disturbance during the daytime (e.g., human intrusion resulting from tourism or recreational caving, guano mining, or entry of goats) (Palmeirim et al. 2005, p. 3; Tarburton 2002, p. 106). Sheath-tailed bat roost sites with chronic disturbance have been observed to be prone to decline and extirpation (Scanlon et al. 2014, p. 453). Therefore, where cave disturbance is an issue, protection of caves from casual entry (public outreach, management, or gating) should be considered to help maintain viability of existing roosts or aid reestablishment of roosts no longer occupied. Within each management unit, both occupied and unoccupied caves and other roost sites should be assessed for their ability to support bat populations, and the highest- priority sites for protection should be identified.

Roosting caves are often in coastal locations at low elevations; therefore they can be catastrophically affected by cyclones (inundation by storm surge and debris), as occurred at the major Tutuila roost with Cyclones Ofa and Val in 1990 and 1991 (Grant et al. 1994, p. 135).

Cyclones can also cause direct storm-related mortality to peʻapeʻa vai individuals (Grant et al. 1994, p. 135). With climate change, cyclones in this region of the Pacific are projected to decrease in frequency but increase in severity (ABOM and CSIRO 2011, p. 195; PCCSP 2014, p. 6). While occasional cyclones are a natural occurrence, their impacts can exacerbate the vulnerability of small or isolated populations. Thus, re-establishment of peʻapeʻa vai populations on multiple islands within its range would be beneficial to increase the overall redundancy of the species and limit its vulnerability to catastrophic events.

Moreover, progressive sea-level rise is expected to continue (ABOM and CSIRO 2011, p. 196- 197; PCCSP 2014, p. 7) and thus is likely to gradually result in routine flooding of the lowest- elevation caves and increase storm surge vulnerability. Both chronic and acute flooding impacts are therefore likely to increase over time. Although root causes are not susceptible to local management, protection of less vulnerable or higher-elevation caves may help to mitigate flooding risk.

There is no specific evidence that disease has contributed to the decline of peʻapeʻa vai populations, but this has not been excluded as a potential factor (Malotaux 2012, pp. 5-6; Palmeirim et al. 2007, p. 517); any field work with peʻapeʻa vai should consider appropriate precautionary measures for avoiding disease spread. Pesticide use in the vicinity of peʻapeʻa vai populations has occurred concurrently with population declines, and in some bat species pesticides have had adverse impacts due to secondary poisoning or reducing insect availability, but specific impacts on peʻapeʻa vai have not been determined (Hutson et al. 2001, p. 138; Malotaux 2012, pp. 5-6; Mickleburgh et al. 2002, p. 119; Palmeirim et al. 2005, p. 35; Tarburton 2002, p. 107). Research on pesticide and disease impacts on sheath-tailed bat populations would be relevant to better understanding whether these are significant issues.

It is likely that the interaction of these issues (foraging habitat degradation, roosting disturbance, catastrophic storm events, low reproductive potential) is responsible for the probable extirpation of peʻapeʻa vai over large regions of its range. The subspecies is likely extirpated from the Samoan archipelago, Tonga, and Vanuatu, and probably remains extant only on smaller islands in the Lau group in Fiji (Scanlon et al. 2014, pp. 455-456). Thus the metapopulation dynamics that likely functioned historically to allow dispersal and gene flow among islands and natural recolonization after catastrophic events now appear to be impaired (Palmeirim et al. 2005, p. 36).

Because natural recolonization originating from source populations in Fiji is likely to be slow or nonexistent given their current condition, it is critical to conduct surveys elsewhere in the range in likely habitat and in formerly known localities where recent occupation is unconfirmed, in order to identify and protect any remnant populations. In American Sāmoa, we recommend visual and acoustic surveys around caves and foraging areas on Taʻu, which have not been fully assessed, and further monitoring of formerly occupied caves on Tutuila (particularly the caves at Anapeʻapeʻa Cove, historically the largest roost in American Sāmoa) for any reoccupation.

If additional populations are not found outside of Fiji, translocation may be considered in order to increase the redundancy of populations. Significant considerations include the ability of a Fiji source population to sustain removal of translocated individuals without adverse effects; suitability of roosting and foraging habitat to support translocated individuals; protocols and

logistical support for safe capture, transport, and release; and international coordination and permitting to transport bats. Previous attempts at translocation of other species of bats (Ruffell and Parsons 2009, Ruffell et al. 2009) have been few and poorly successful due to released bats attempting to return to their point of origin, so any translocation methods should be carefully evaluated and tested in less vulnerable species of insectivorous bats. Development of any translocation plan should take into account the general guidelines and considerations for translocations summarized by IUCN (2013).

Protection of remaining populations in Fiji is particularly critical (Scanlon et al. 2014, pp. 455- 456). Continuation of the ongoing efforts of the Fiji Bat Conservation Initiative to survey bat populations throughout Fiji, monitor new and historical roosts, increase public awareness of bat conservation, and work with government and local partners to support targeted roost protection and adjacent land management is necessary to stabilize the population in Fiji and support eventual reestablishment in other management units. Research useful to inform recovery implementation would include assessment of habitat and microclimate preferences for roosting sites and/or foraging locations. NatureFiji is currently finalizing its National Bat Conservation Strategy (N. Thomas-Moko in litt. 2020), which should inform the prioritization of sites for protection.

Recovery Strategy for Maʻomaʻo Our objective is to manage predators and habitat to establish self-sustaining, reproductively viable maʻomaʻo populations and increase population redundancy within Sāmoa and American Sāmoa. The most significant threats include predation by rats and cats, and loss and fragmentation of mature forest habitat.

Maʻomaʻo life history, in comparison to other honeycreeper species, is characterized by a slow rate of reproduction and an extended incubation period (Stirnemann et al. 2015, pp. 373-378). Thus the species is particularly vulnerable to nest failure and predation. A recorded nighttime nest predation event and attack on an incubating female by a rat, as well as presence of rats near occupied nests and disappearance of females during the breeding season, indicate that rat predation is a mortality factor (Butler and Stirnemann 2013, p. 28). Cats occur in montane forest habitat occupied by maʻomaʻo (Atherton and Jefferies 2012, pp. 101-106); because maʻomaʻo fledglings are poor flyers and spend time on the ground for weeks after fledging, they are likely to be particularly vulnerable to cat predation (Butler and Stirnemann 2013, p. 31). Thus we expect nest predation by rats and predation on fledglings and/or adults by cats to be significant sources of mortality (Butler and Stirnemann 2013, pp. 28-31). Therefore, management to control rat and cat populations (deployment of self-setting traps and/or rodenticide application, etc.) within the current range of maʻomaʻo or in potential reintroduction locations is a high priority. On smaller islets complete eradication of rats may be feasible; on the larger islands, ongoing control programs are likely to be necessary to reduce rat abundance. Predator control would be an important component of management either to support viability of existing populations in Sāmoa or to prepare potential translocation sites and maintain translocated populations.

Maʻomaʻo distribution is associated with mature forest habitat (tall trees are used for song perches, nest sites, and foraging) (Butler and Stirnemann 2013, p. 30; Engbring and Ramsey 1989, p. 68). A high diversity of tree species with a variety of seasonal flowering periods also

likely helps to ensure continuous nectar availability for foraging (Butler and Stirnemann 2013, p. 30). Therefore, forest clearance due to development or agriculture renders habitat unsuitable for the species. Invasive plant species have potential to affect maʻomaʻo by altering forest structure and reducing vegetation diversity (ASCC 2010, pp. 13-15, 22; Atherton and Jefferies 2012, pp. 101-106; Butler and Stirnemann 2013, p. 22; Space and Flynn 2002, entire). Pigs are widespread in Samoan forests and can also change habitat structure and clear vegetation by rooting and wallowing, as well as further promoting invasion by nonnative plants (ASCC 2010, p. 15; Atherton and Jefferies 2012, pp. 17, 101, Space and Flynn 2002, pp. 8-10). Therefore protection of diverse mature forests (managing sites to retain intact forest canopy, locating and controlling invasive plants, and controlling pig populations) is important to maintain suitable maʻomaʻo habitat.

Catastrophic cyclone impacts can damage or destroy forest canopy structure, as well as causing direct storm-related mortality to birds (Butler and Stirnemann 2013, p. 41; MNRE 2006, pp. 5-7; Wiley and Wunderle 1993, entire). With climate change, cyclones in this region of the Pacific are projected to decrease in frequency but increase in severity (ABOM and CSIRO 2011, p. 195; PCCSP 2014, p. 6). While occasional cyclones are a natural occurrence, their impacts can exacerbate the vulnerability of small or isolated populations. Thus, establishment of maʻomaʻo populations on other islands would be beneficial to increase the overall redundancy of the species and limit its vulnerability to catastrophic events.

Currently maʻomaʻo populations exist only on the islands of ʻUpolu and Savaiʻi in Sāmoa (MNRE 2006, p. 1). Options for translocation sites should be evaluated (including islets in Sāmoa such as Nuʻutele and Nuʻulua, the previously occupied island of Tutuila in American Sāmoa, and other American Sāmoa islands including Ofu, Olosega, and Taʻu that are outside the species’ known historical range). Translocation considerations include: the ability of source population(s) in Sāmoa to sustain removal of translocated individuals; potential for captive propagation; identifying potential founder individuals and assessing genetic representation; vegetation characteristics, habitat suitability, and available habitat area at translocation sites; presence of predators and logistical feasibility of control; potential for competitive interactions with other birds (e.g., wattled honeyeater [Foulehaio carunculatus]); protocols for safe capture, transport, and release; review and assessment of methodology and outcome for any translocation efforts attempted with similar species; and international coordination and permitting to transport maʻomaʻo individuals. Development of any translocation plan should take into account the general guidelines and considerations for translocations summarized by IUCN (2013).

A more detailed assessment of maʻomaʻo territory sizes and their associations with forest species composition should be useful to anticipate the size of the population that various translocation sites could support and their overall habitat suitability. Note that estimated territory sizes of comparable species range widely from 1 to 10 hectares (2.5 to 25 acres) (giant forest honeyeater [Gymnomyza viridis] see MNRE 2006; Jackson and Jit 2007, p. 107) to 50 to 75 hectares (124 to 185 acres) (crow honeyeater [Gymnomyza aubreyana], see Higgins et al. 2008, p. 671). Thus the species-specific territory size for maʻomaʻo needs to be more clearly understood, because it has large consequences for the amount of land area that will need to be actively managed to meet target population sizes for the species. It is unlikely that islets would support more than a few breeding pairs, but if predators could be locally eradicated these sites might still act as a

15 supporting source population. Introduction to larger islands would improve overall redundancy of the species against catastrophic events, but would likely need to be supported by ongoing predator control efforts. In some cases predator control efforts for other listed species (particularly on Ofu and Olosega - see discussion under tuʻaimeo recovery strategy below) might be coordinated with maʻomaʻo translocation.

Sāmoa earlier completed a recovery plan for the maʻomaʻo (MNRE 2006), that established an overall recovery goal and objectives for a 10-year period (2006 to 2016); identified several target areas for conservation; and discussed issues relevant to implementing monitoring and management. The recovery objectives included: • Conserve and manage key forest areas on ʻUpolu and Savaiʻi to secure maʻomaʻo populations on both islands. • Carry out detailed surveys to identify the numbers of pairs at certain sites and establish monitoring to record any changes in this over time. • Increase the understanding of the breeding and feeding ecology of the maʻomaʻo to aid species recovery. • Establish populations on rat-free islands or new mainland sites. • Evaluate the development of a captive management programme. • Develop partnerships to assist in the recovery of the maʻomaʻo through provision of funds, support, or expertise. • Establish a Threatened Bird Recovery Group to oversee the implementation and review of this plan and those of other priority bird species.

We concur with this overall strategy as described in the 2006 MNRE recovery plan. Since publication of the recovery plan, some additional research on maʻomaʻo biology and breeding ecology has been done (Butler and Stirnemann 2013, Stirnemann et al. 2015), and rat eradication was attempted on the islets of Nuʻutele and Nuʻulua (A. Tye in litt. 2012) but no captive propagation or translocation of populations to new sites has yet occurred. As of 2016, the plan is currently due to be reviewed for an update; if revisited the plan could be revised as appropriate based on the current results of research, monitoring, and management implementation data.

Recovery Strategy for Tuʻaimeo Our objective is to manage predators and habitat to establish self-sustaining, reproductively viable tuʻaimeo populations and increase population redundancy within American Sāmoa. The most significant threat is predation by rats and cats.

Because the tuʻaimeo forages primarily on the ground and in the forest understory and builds its nest from 1 to 3 meters (3 to 10 feet) off the ground, it is vulnerable to mammalian predators (Baptisa et al. 1997, pp. 180-186; Kayano et al. 2019, p. 19). Rats and cats occur throughout American Sāmoa, including on Olosega and Ofu (Craig 2009, pp. 15-18), and one observed tuaimeo mortality was likely a result of cat predation (Kayano et al. 2019, pp. 19, 23-24). Nest predation by rats and predation by cats on all life stages (nests, fledglings, and adults) are likely to be significant mortality factors (Steadman and Freifeld 1998, p. 617; Watling 2001, p. 118). Its small clutch size (1-2 eggs) increases its vulnerability to this threat (Baptisa et al. 1997, pp. 180-186). Therefore, a predator control program on Ofu and Olosega will be a necessary component of improving survivorship and reproductive success. A control program to reduce or

eliminate the feral cat population, accompanied by deployment of self-setting traps and/or rodenticide to reduce rat abundance, would be expected to substantially reduce predation pressure and allow an increase in tuʻaimeo population.

Habitat used by tuʻaimeo in American Sāmoa includes steep, forested slopes with an open understory, as well as coastal areas (Kayano et al. 2019, pp. 13, 18-19). Thus, forest clearing could potentially reduce habitat suitability, although the rugged terrain on Ofu and Olosega has historically made this a minor issue. Because of the species’ association with forest understory for foraging and breeding, invasive plants have potential to affect habitat suitability by altering understory structure and reducing vegetation diversity (ASCC 2010, pp. 13-15, 22; Space and Flynn 2002, entire). Pigs can also alter habitat structure and clear vegetation by rooting and wallowing, and can promote invasion by nonnative plants (ASCC 2010, p. 15; Space and Flynn 2002, pp. 8-10). Therefore, protection of forest habitat (managing sites to retain forest canopy, locating and controlling invasive plants, and controlling pig populations) is important to maintain suitable tuʻaimeo habitat.

Additional research on habitat association; genetic assessment and comparison with populations of the species outside of American Sāmoa (Sāmoa, Fiji, Tonga, etc.); demographic assessment of survivorship, lifespan, and reproductive success; and home range size will be helpful to inform management. Improvement in survey methods to develop a more accurate assessment of population size is important for tracking the progress of recovery (Kayano et al. 2019, pp. 22- 24).

Because tuʻaimeo in the American Sāmoa DPS occur only on Ofu and Olosega, coordination and outreach with villages on those islands is critical to maintain community awareness and support for these management actions.

Catastrophic cyclone impacts can damage or destroy forest canopy structure, as well as causing direct storm-related mortality to individual birds (Wiley and Wunderle 1993, entire). With climate change, cyclones in this region of the Pacific are projected to decrease in frequency but increase in severity (ABOM and CSIRO 2011, p. 195; PCCSP 2014, p. 6). While occasional cyclones are a natural occurrence, their impacts can exacerbate the vulnerability of small or isolated populations. Ofu and Olosega are immediately adjacent to one another and any cyclone impacts are likely to affect both of these islands similarly. Thus, establishment of tuʻaimeo populations on other islands within American Sāmoa (i.e., Taʻu and/or Tutuila) would be beneficial to increase the overall redundancy of the species and limit its vulnerability to catastrophic events. Habitat on Taʻu is similar to that on Ofu and Olosega (ASCC 2010, p. 7) and is likely to be suitable. Rat and cat predation would similarly be an issue for translocated populations on Taʻu or Tutuila, so rat and cat control programs would be helpful to support any translocation effort. Feasibility and logistical considerations for predator control likely differ on Tutuila; the island is larger and more readily accessible than Taʻu, but also has more extensively disturbed forest and is more heavily populated (ASCC 2010, p. 7). In addition, Taʻu supports a native population of Pacific tree boa (Candoia bibroni) (Craig 2009, p. 60), which might also prey on tuʻaimeo and should be considered when evaluating the island’s suitability as a translocation site.

Recovery Strategy for Eua zebrina and Ostodes strigatus Our objective is to manage predators and habitat to improve viability of Eua zebrina (and, if relocated, Ostodes strigatus) populations and increase population redundancy at suitable localities within American Sāmoa. The most significant threat to both species is predation by nonnative invertebrates.

Because partulid snails have slow growth rates and relatively low reproductive capacity, and evolved in the absence of significant predation, they are particularly vulnerable to predator introductions (Cowie 2001, p. 218). As a result, numerous partulid species on various Pacific islands have historically suffered population declines and extinction due to introduction of predatory snails (particularly Euglandina) and the New Guinea flatworm (Platydemus manokwari), as well as rats (Cowie 1992, pp. 170-175). Mortality from nonnative predators is thus a primary concern for Eua zebrina (as well as the nonlisted American Sāmoa species of Samoana).

Ostodes strigatus is not closely related to the partulid snails, and its life history is not well known. However, like Eua zebrina, it evolved in the absence of predators, and its ground- dwelling habit may also have increased its vulnerability to predation (Girardi 1978, p. 224). The species’ rapid decline within a few years of the introduction of Euglandina to Tutuila, coupled with the observation of abundant Euglandina at sites recently occupied by Ostodes strigatus, is strongly suggestive that predation was responsible for its decline (Miller 1993, pp. 24-27).

Planning for management of these two listed snails is complicated by the lack of current information about their distribution and status of their threats. As noted above, Eua zebrina was observed in 1998 surveys on Tutuila and Ofu (Cowie and Cook 1999, p. 30-31); the Tutuila population remains extant but has not been systematically surveyed to assess its distribution or abundance since that time (A. Miles in litt. 2015). No more recent information on the Ofu population is available. Ostodes strigatus has not been observed since 1992 and may be extinct, but systematic surveys across its range in the western part of Tutuila are needed (Cowie and Cook 1999, p. 24; Miller 1993, pp. 11, 24-27).

Euglandina has been established on Tutuila since the 1980s (Eldredge 1988, pp. 122-125), was widely distributed on the island in the 1992 and 1998 surveys, and continues to be regularly observed there (Miller 1993, pp. 10, 23-28; Cowie and Cook 1999, pp. 36, 71-85). On Taʻu its current distribution or establishment is unclear; one individual was observed in the 1992 survey, but none were observed in 1998 (Miller 1993, p. 10; Cowie and Cook 1999, pp. 87-91). Euglandina has never been reported from Ofu or Olosega, or from Nuʻusetoga islet near Tutuila (Cowie and Cook 1999, pp. 93-98; Cowie et al. 2003, p. 39); however, it is unknown whether it may have arrived on these islands since the last surveys in the 1990s. Other species of introduced carnivorous snails have also been observed on several islands (Cowie and Cook 1999, pp. 20, 36) while these are generally smaller and likely to be a less significant threat to native snails than is Euglandina, they may have some adverse effect. Platydemus manokwari is reported to have been detected on Tutuila and Taʻu as of 2004, but no further information is available and it is not clear whether or how widely it is currently established (Craig 2009, p. 84). Rats, though probably a less significant predator, are widely distributed throughout the American Sāmoa islands (Craig 2009, p. 60).

Consequently, a systematic survey to assess the distribution and abundance of the listed snails as well as their predators is critical to planning next steps for recovery. Repeated surveys over time using comparable methods will be necessary to assess ongoing population trends and evaluate the success of recovery actions. Understanding whether Ofu and Olosega still remain free of Euglandina and Platydemus manokwari, and clarifying the status of these predators on Taʻu, is crucial to planning any translocation measures. Although Eua zebrina appears so far to be persisting on Tutuila in the presence of Euglandina and perhaps P. manokwari (Cowie and Cook 1999, pp. 36, 71-85; Craig 2009, p. 84; A. Miles in litt. 2015), in order to assess whether its population is viable more information is needed about their interactions, comparative distribution patterns, demography, and population trends. No effective recovery actions can be implemented for Ostodes strigatus unless a population is rediscovered.

To date, no effective methods are available for controlling or eradicating Euglandina or Platydemus manokwari on islands where they are established. In Hawaiʻi, Euglandina has been successfully removed from small exclosures, but only by repeated and extremely labor-intensive hand-removal supported by snail-proof fencing (Rohrer et al. 2013, pp. 17-18). Emerging genetic technologies (McLaughlin and Dearden 2019, pp. 5-6) could conceivably provide more effective tools for eradication of introduced Euglandina or P. manokwari populations, and could have broad application on islands around the Pacific. However, no such applications have yet been developed. If feasible, these applications would need to be designed specifically for the particular genetics and reproductive biology of the target species, tested for effectiveness, and implemented only with full public outreach and informed local agreement.

Rat control could also benefit native snail populations, and in some cases could be coordinated with management for other listed species. Rats are widespread throughout the American Sāmoa islands (Craig 2009, p. 60). On the larger islands (Tutuila, Ofu, Olosega, and Taʻu) rat control programs (deployment of self-setting traps, and/or rodenticide) would require ongoing maintenance effort but could reduce rat populations and improve snail survival within target areas. On smaller islets full eradication may be feasible. In particular, rat eradication could provide substantial benefits on Nuʻusetoga, where an isolated Eua zebrina population was found in 1992 (Miller 1993, p. 13) with evidence of rat predation but no invertebrate predators, and such measures should be considered there.

Because both of these snails occur in native forest habitat, forest protection measures including managing sites to retain forest canopy, locating and controlling invasive plants, and controlling pig populations are likely to benefit habitat suitability (Cowie 2001, p. 219). Because snails can be sensitive to microclimatic conditions, including temperature regime, humidity, and precipitation (Cowie 1992, p. 176; Clarke 2019, p. 24), further modeling and assessment of how these parameters are likely to be affected by forest structure, vegetation type, and climate change would be helpful for evaluating current and prospective habitat suitability of localities.

Depending on new survey results and the current distribution of native snails and introduced predators across the American Sāmoa islands, it may be appropriate to increase population redundancy by translocating native snails to islands where they do not currently exist, or to establish captive populations. The most recent available information, which requires updating,

suggests that the islands of Ofu and Olosega, and the islet Nuʻusetoga, may remain free of invertebrate predators and be able to support translocated native snail populations. The status of predators on Taʻu is unclear. In addition to predation considerations, a habitat evaluation of potential translocation sites (e.g., forest structure, vegetation composition, disturbance, presence of invasive plants, precipitation and microclimate) should be done to assess where introduction is most likely to be successful. Development of any translocation plan should take into account the general guidelines and considerations for translocations summarized by IUCN (2013). A genetic and morphological assessment of Eua zebrina populations native to different localities within Tutuila and Ofu would be helpful to assess whether they constitute distinct management subunits that should be individually conserved. If Ostodes strigatus individuals are rediscovered on Tutuila, it is probable that their population will be at high risk from predation and translocating them to a more viable site should be urgently considered.

Alternatively, and assuming greater importance if surveys show the Eua zebrina population is in decline and Euglandina and/or Platydemus manokwari are found to exist on islands where they have not previously been detected, individuals of Eua zebrina or (if new populations are located) Ostodes strigatus could be brought into controlled propagation programs. Controlled propagation, whether at new facilities within American Sāmoa or in coordination with snail propagation programs elsewhere, should be one component of a larger strategy to reintroduce native snails into natural habitat. Best practices for captive management and propagation of Partula snails have already been developed (Clarke 2019, entire) and should be considered and applied appropriately for related partulid species with the advice and participation of experienced individuals.

In addition, because similar conservation issues apply to other non-listed native snails in American Sāmoa, including but not limited to the Samoana species and Ostodes adjunctus (Cowie 1998, pp. 25, 53-55), survey information and translocation or controlled propagation efforts should be concurrently considered for these species as well, both for their conservation and to help avoid the need for future listing.

Preventing introduction of Euglandina or Platydemus manokwari to new islands within American Sāmoa is critical to maintaining suitable habitat where viable populations of the listed snails can exist. Introductions could occur accidentally in the course of moving passengers or cargo between islands, particularly if vegetation or soil are transported. Predators could also be introduced deliberately. This is most likely to occur in an attempt to control pest populations of the giant African snail (Achatina fulica), although there is little or no evidence of biocontrol having been successful for this species (e.g., Eldredge 1988, pp. 119-125). Consequently, preventing introduction of A. fulica to new islands is also vital for native snail recovery in order to avoid incentives for predator introduction. Appropriate measures should involve outreach to raise public awareness, coordination with ports and shipping facilities, and development and implementation of a territorial government biosecurity policy. Shipping from outside American Sāmoa should also be monitored to avoid further introduction of new predators; in particular, the second Euglandina clade already present in Hawaiʻi could well exacerbate existing predation impacts if it were to become established in American Sāmoa (Meyer et al. 2017a, p. 1402).

B. RECOVERY CRITERIA

Section 4(f)(1)(B)(ii) of the Act states that each recovery plan shall incorporate, to the maximum extent practicable, “objective, measurable criteria which, when met, would result in a determination... that the species be removed from the List.” Legal challenges to recovery plans (see Fund for Animals v. Babbitt, 903 F. Supp. 996 (D.D.C. 1995)) and a Government Accountability Audit (GAO 2006) also have affirmed the need to frame recovery criteria in terms of threats assessed under the five listing factors.

Recovery criteria serve as objective, measurable guidelines to assist in determining when an endangered species has recovered to the point that it may be downlisted to threatened, or that the protections afforded by the Act are no longer necessary and the species may be delisted. Delisting is the removal of a species from the Federal Lists of Endangered and Threatened Wildlife and Plants (Lists). Downlisting is the reclassification of a species from endangered to threatened. The term “endangered species” means any species (species, subspecies, or DPS) that is in danger of extinction throughout all or a significant portion of its range. The term “threatened species” means any species that is likely to become an endangered species within the foreseeable future throughout all or a significant portion of its range.

Recovery criteria represent our best assessment, at the time the recovery plan is completed, of the conditions that would likely result in a determination that listing under the Act as threatened or endangered is no longer required. However, revisions to the Lists, including delisting or downlisting a species, must reflect determinations made in accordance with sections 4(a)(1) and 4(b) of the Act. Section 4(a)(1) requires that the Secretary determine whether a species is an endangered species or threatened species because of threats to the species, based on an analysis of the five listing factors in section 4(a)(1). Section 4(b) require that the determination be made “solely on the basis of the best scientific and commercial data available.” Thus, while recovery plans provide important guidance to the Service, States, and other partners on methods of minimizing threats to listed species and measurable criteria against which to measure progress towards recovery, they are guidance and not regulatory documents.

Thus, a decision to delist or downlist a species is informed by the recovery criteria but is ultimately based on an analysis of threats using the best scientific and commercial data then available. When changing the status of a species, we first propose the action in the Federal Register to seek public comment and peer review, after which we announce a final decision in the Federal Register.

The species addressed in this recovery plan should be considered for downlisting and delisting when the following objective[s] and criteria have been met. Downlisting and delisting criteria are subject to change as additional information becomes available about species biology and threats.

1. Peʻapeʻa Vai

Objective – Manage threats to enhance and reestablish self-sustaining, reproductively viable populations of peʻapeʻa vai in multiple island groups across its historical range.

Downlisting Criteria: To downlist the peʻapeʻa vai from endangered to threatened, the following criteria must be met:

(1) Peʻapeʻa vai populations are established and stable in Fiji and at least 1 other management unit (either American Sāmoa or Sāmoa) over a period of at least 10 years, with active, consistently occupied roosts represented on multiple islands in each management unit. Peʻapeʻa vai population data in each management unit show a stable or positive growth curve over this time period, as determined through quantitative surveys of abundance or an index of abundance derived from quantitative surveys or demographic monitoring. The average population over this time period is estimated to be at least 5,000 individuals in each management unit.

Downlisting Criteria Rationale: Given a generation time of approximately 2.5 years (Pacifici et al. 2013, entire; USFWS 2020a, p. 6), monitoring over a 10-year period should allow tracking of population size over several generations to assess the sustained population trend. Protection and improvement of population status in Fiji, the only management unit where populations are known to be extant, is a necessary precondition for restoring populations in other island chains. If populations also successfully become reestablished within one other management unit this would increase redundancy to catastrophic events such as cyclones, and would provide evidence that the declining trend has been partially reversed such that the overall population status of the subspecies is beginning to be ameliorated. Establishment of populations in additional management units will increase the ability of the species to function as a metapopulation with natural dispersal and recolonization.

While information about historical population sizes within management units varies, observations from American Sāmoa, Sāmoa, and Fiji suggest that larger roosts historically included several thousand individuals and that overall populations within an island chain likely ranged to tens of thousands of individuals (Amerson et al 1982, p. 74; Ollier et al. 1979, Palmeirim et al. 2005, pp. 31-32). (Historical populations in Tonga and Vanuatu are less well known and the habitat in these management units may be insufficient to support comparable population levels.) The most recent estimates of roost sizes in Fiji indicated that the population had contracted in distribution and declined in numbers by 80 percent over the previous decade from an already depressed population of several hundred individuals (Scanlon et al. 2014, p. 453). A sustained population level of

22 approximately 5,000 within a management unit would represent a substantial improvement from current conditions but would likely remain well below historical population sizes and carrying capacity. This population threshold is a preliminary estimate and a population viability analysis for the species could be informative to adjust this threshold and project probability of persistence over a given time period.

Protection of roost sites from disturbance and maintenance of forest habitat in the vicinity should help to address major factors involved in population decline so as to keep existing roosts viable; even if roost sites are abandoned and not currently active, protection measures would still be valuable to encourage recolonization. Sites identified in Fiji appear to be the highest priority for immediate protection (and given physiographic constraints on presence of roost sites, might not be readily substituted) but do not constitute a complete list; sites to be protected in other management units should also be locally identified and prioritized. At present the relative role of various threats (e.g., pesticides, disease, habitat modification, climate change) in the rangewide decline of peʻapeʻa vai remains unclear, but targeted research and monitoring should improve our understanding of the magnitude and mechanism of these threats and their impacts on population viability. If population size and distribution improves substantially we expect that threat severity will have been correspondingly ameliorated.

Delisting Criteria: To delist the peʻapeʻa vai, the following criteria must be met:

Delisting Criteria Rationale: Similarly to the discussion above under Downlisting Criteria, monitoring of population status and trend over an additional 10 years would allow tracking of population status and trend over several additional generations.

If viable populations are established across three management units, metapopulation structure and function will have been established across a majority of the species’ historical range, providing redundancy against cyclones or other catastrophic events and opportunity for natural patterns of dispersal and recolonization. A sustained population level of approximately 10,000 within a management unit would represent a substantial improvement from the downlisting threshold and be indicative of overall viability of the population within the management unit. In some cases we expect this would still remain somewhat below historical population sizes and carrying capacity (the American Sāmoa population size was estimated at 11,000 by Amerson [1982], most concentrated in a single roost, but other management units such as Fiji and Sāmoa that covered more geographic area, islands, and/or potential roost sites plausibly had larger populations). Historically, populations in the Vanuatu and Tonga management units are less well known and these management units may be less able to support large populations, so management units in Fiji, Sāmoa, or American Sāmoa may more easily meet this threshold. This population threshold is a preliminary estimate and a population viability analysis for the species could be informative to adjust this threshold and project probability of persistence over a given time period.

As discussed above, availability of suitable foraging habitat and protection of roosts from disturbance, along with adequate management of other threats to the species, are necessary for recovering management unit populations. We expect that substantial improvement in population size and distribution will correspond with amelioration of overall threat severity.

2. Maʻomaʻo

Objective – Manage predators and habitat to establish self-sustaining, reproductively viable maʻomaʻo populations and increase population redundancy within Sāmoa and American Sāmoa.

Downlisting Criteria: To downlist the maʻomaʻo from endangered to threatened, the following criteria must be met:

(2) Mature forest habitat in Sāmoa is protected with long-term management commitments and habitat restoration has been completed to the extent that the amount of habitat suitable for maʻomaʻo is sufficient to sustain population targets identified in Downlisting Criterion 1.

24 (3) Predator control programs in maʻomaʻo habitat on Savaiʻi and ʻUpolu are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Downlisting Criterion 1 are met.

Downlisting Criteria Rationale: Typical generation time of maʻomaʻo is uncertain, but as maximum lifespan of other honeyeater species ranges from 7 to 15 years, and the maʻomaʻo has a comparatively large body mass relative to other species in the same family, we expect that generation time is relatively long (Geering 2002; USFWS 2020b, pp. 17-18). Thus we have identified a population monitoring time frame of 20 years to allow tracking of population size over several generations and assessment of sustained population trend.

The maʻomaʻo population across Savaiʻi and ʻUpolu in Sāmoa was roughly estimated at 500 individuals in 2006 (MNRE 2006, p. 4; Birdlife International 2016); this likely represented an already depressed population following observed range contraction in the 1990s, historic forest fragmentation, and several cyclones. Documenting a Sāmoa population of approximately 1,000 individuals encompassing both major islands, with a stable to increasing population trend, would indicate a substantial shift toward improved conservation status. This population threshold is a preliminary estimate and a population viability analysis for the species, combined with improved assessment of maʻomaʻo habitat requirements and territory sizes, could be informative to adjust this threshold and project probability of persistence over a given time period.

Given the known threats to the species, we expect that achieving such improvements in population abundance and trend will require active management of predator populations and protection and restoration of mature forest habitat. Assessment that improved status is secure would also require ongoing long-term commitments to continuation of such management. We expect that substantial improvement in population size and distribution will correspond with amelioration of overall threat severity.

Delisting Criteria: To delist the maʻomaʻo, the following criteria must be met:

(1) Over a period of an additional 20 years, maʻomaʻo population data show a stable or increasing trend (i.e., finite rate of population increase, or lambda, greater than or equal to 1) that is statistically significant, as determined through quantitative surveys of abundance, or an index of abundance derived from quantitative surveys or demographic monitoring. The average population throughout that time period is estimated to be at least 1,000 individuals on both Savaiʻi and ʻUpolu; or the average population is estimated to be at least 750 individuals on both Savaiʻi and ʻUpolu, and 1 or more self-sustaining populations totaling at least 250 individuals are established within American Sāmoa.

(2) Mature forest habitat in Sāmoa and at any translocation localities in American Sāmoa is protected and restoration has been completed to the extent that the amount of habitat suitable for maʻomaʻo is sufficient to sustain population targets identified in Delisting Criterion 1.

(3) Predator control programs in maʻomaʻo habitat on Savaiʻi, ʻUpolu, and any translocation localities in American Sāmoa are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Delisting Criterion 1 are met.

Delisting Criteria Rationale: Similarly to the discussion under Downlisting Criteria above, monitoring of maʻomaʻo population status and trend over an additional 20 years would allow tracking of population status and trend over several additional generations. After downlisting criteria are met, documenting a Sāmoa population continuing to increase to approximately 1,500 individuals encompassing both major islands, with a stable to increasing population trend, in combination with establishment of an American Sāmoa population, would indicate a substantial shift toward improved conservation status and indicate the species is likely secure. Reestablishment of maʻomaʻo within American Sāmoa would substantially improve the redundancy of the species against cyclone impacts or other catastrophes. However, the extent of available habitat within American Sāmoa is less than on Savaiʻi and ʻUpolu (MNRE 2006, p. 8), so it is likely that the overall achievable carrying capacity is lower and translocated populations will be smaller; thus we provide a smaller population size target for American Sāmoa. If reestablishment of the species in American Sāmoa fails or is otherwise infeasible, documentation of a thriving population in Sāmoa could alternatively provide evidence of full recovery, but we expect that population levels would need to be correspondingly higher to compensate for the lack of an American Sāmoa population and lower redundancy of the species. These population thresholds are a preliminary estimate and a population viability analysis for the species, combined with improved assessment of maʻomaʻo habitat requirements and territory sizes, could be informative to adjust this threshold and project probability of persistence over a given time period.

Given the known threats to the species, we expect that achieving such improvements in population abundance and trend and/or establishment of translocated populations will require active management of predator populations and protection and restoration of mature forest habitat. Assessment that improved status is secure would also require ongoing long-term commitments to continuation of such management. We expect that substantial improvement in population size and distribution will correspond with amelioration of overall threat severity.

3. Tuʻaimeo

Objective – Manage predators and habitat to establish self-sustaining, reproductively viable tuʻaimeo populations and increase population redundancy within American Sāmoa.

Downlisting Criteria: To downlist the tuʻaimeo from endangered to threatened, the following criteria must be met:

(1) Over a period of at least 10 years, tuʻaimeo population data show a stable or increasing trend (i.e., finite rate of population increase, or lambda, greater than or equal to 1) that is statistically significant, as determined through quantitative surveys of abundance, or an index of abundance derived from quantitative surveys or demographic monitoring; and the average population throughout that time period is estimated to be at least 250 individuals on at least 1 island (Ofu or Olosega).

(2) Forest habitat on Ofu and Olosega is protected and managed such that the amount of habitat suitable for tuʻaimeo is sufficient to sustain population targets identified in Downlisting Criterion 1.

(3) Predator control programs on Ofu and Olosega are in effect with long-term commitments, and are effectively managing predation mortality to the extent that population targets in Downlisting Criterion 1 are met.

Downlisting Criteria Rationale: No information is available about generation time or lifespan of tuʻaimeo or closely related Pacific island dove species. By comparison, typical annual mortality in the wild of adult mourning doves (Zenaida macroura), which have similar body mass although they exist in a very different ecological setting, is approximately 55 percent (Otis et al. 2008). In absence of species-specific information, we have identified a population monitoring time frame of 10 years, which should allow tracking of population size over several generations and assessment of sustained population trend. This timeframe may be reassessed if more species-specific life history information becomes available.

Tuʻaimeo appear to be difficult to survey accurately, and historical population estimates allow only a general assessment of population size. Amerson (1982) estimated a population size of about 100 birds on Ofu and possibly Olosega. Using habitat-specific densities derived from area searches, Kayano et al. (2019) estimated a population size of 145 on Ofu and 104 on Olosega; however, they were unable to calculate statistical confidence and considered the estimate extremely tentative. Given the ongoing threat of mortality from rat predation on nests and cat predation on adults and fledglings, we expect that these estimates reflect population levels that are depressed relative to historical numbers and potential carrying capacity. The threshold of 250 individuals on 1 island thus represents a level that would likely indicate substantial improvement in population status; however, this threshold may be revisited if improved survey information or a population viability analysis become available.

such management. We expect that substantial improvement in population size and distribution will correspond with amelioration of overall threat severity.

Delisting Criteria: To delist the tuʻaimeo, the following criteria must be met:

(2) Forest habitat on Ofu and Olosega and any translocation localities is protected and managed such that the amount of habitat suitable for tuʻaimeo is sufficient to sustain population targets identified in Delisting Criterion 1.

Delisting Criteria Rationale: Similarly to the discussion under Downlisting Criteria above, monitoring of tuʻaimeo population status and trend over an additional 10 years would allow tracking of population status and trend over several additional generations. After downlisting criteria are met, documenting a population continuing to increase to approximately 500 individuals on at least 2 islands, with a stable to increasing population trend, in combination with establishment of an American Sāmoa population, would indicate a substantial shift toward improved conservation status and indicate the species is likely secure. While expanded populations on both Ofu and Olosega would suffice to meet this threshold, establishment of a population on Tutuila or Taʻu could provide an alternative pathway to meet this target and would improve the overall redundancy of the species to catastrophic events such as cyclones.

Given the known threats to the species, we expect that achieving such improvements in population abundance and trend and/or establishment of translocated populations will require active management of predator populations and the continued availability of protected mature forest habitat. Assessment that improved status is secure would also require ongoing long-term commitments to continuation of such management. We expect that substantial improvement in population size and distribution will correspond with amelioration of overall threat severity.

4. Eua zebrina and Ostodes strigatus

Objective – Manage predators and habitat to improve viability of Eua zebrina or Ostodes strigatus populations and increase population redundancy at suitable localities within American Sāmoa.

Downlisting Criteria: To downlist Eua zebrina or Ostodes strigatus from endangered to threatened, the following criteria must be met for the species:

(1) At least 6 stable populations (possibly actively managed) exist on 1 or more islands within American Sāmoa. To be considered stable, a population must number at least 500 individuals distributed across all size classes combined, and must have a population growth curve or index trend that is stable or positive for at least 4 of 5 sequential years. If multiple management subunits have been identified for the species based on morphological or genetic characters, each subunit must comprise one or more of these stable populations.

(2) Each population in Downlisting Criterion 1 occurs on suitable habitat that is managed to protect native forest vegetation.

Downlisting Criteria Rationale: For additional discussion of issues relevant to recovery criteria for these snail species, see also the Amendment to the Recovery Plan for Oahu Tree Snails of the Genus Achatinella (USFWS 2019c), which addresses Hawaiian snail species that are generally similar to Eua zebrina in their ecology and vulnerability to predation. However, by comparison to Achatinella, partulid species tend to have shorter lifespan, earlier reproductive maturity, and a greater reproductive rate, typically producing 1 offspring every 20 days over a lifespan of 5 years (Cowie 1992, pp. 167-180). A 5-year monitoring period appears appropriate to capture population trends of Eua zebrina spanning several generations. The frequency of cyclones that have major effects on forest habitat is greater in American Sāmoa than on Oahu; future projections of cyclone activity indicate that storm frequency is likely to decrease, but storm severity is likely to increase (ABOM and CSIRO 2011, p. 195; PCCSP 2014, p. 6). Consequently, to allow for sufficient population size for recovery from greater cyclone impacts, we establish a threshold for Eua zebrina that is larger than the 300 individuals identified for Achatinella. Pending a detailed assessment of geographic variation and threats, maintenance of 6 populations of 500 individuals should be sufficient to allow conservation of ecological diversity of the species. Although life history characteristics of Ostodes strigatus are poorly known, we have no information indicating that the species would differ substantially in appropriate

population thresholds or monitoring period; if it is rediscovered, thresholds should be reassessed as appropriate in light of species-specific information.

The primary conservation concern for both snails is risk of predation by nonnative predators (Cowie 1992, pp. 170-175). Presence of suitable forest habitat is also necessary for both species (Cowie 2001, p. 219). While Eua zebrina populations on Tutuila appear at present to be persisting in the presence of Euglandina, its existing population size, distribution, and the demographic effects of its interaction with predators need to be evaluated. If Ostodes strigatus is rediscovered, given its rapid historic decline after Euglandina introduction any persisting population on Tutuila is unlikely to be adequately free of predation risk. Establishment of translocated populations on other islands or at sites actively managed to exclude predators may be necessary to achieve these criteria.

Because Eua zebrina is known from both Tutuila and Ofu, any genetic or morphological distinctions between the island populations (or among populations on Tutuila) should be assessed to identify appropriate management subunits, and considered in management so that the genetic diversity of the species continues to be preserved.

Delisting Criteria: To delist Eua zebrina or Ostodes strigatus, the following criteria must be met for the species:

(2) Each population in Delisting Criterion 1 occurs in suitable habitat that is managed to protect native forest vegetation and is capable of supporting expansion of the occupied range, positive population growth, and establishment of new populations through natural dispersal.

(3) Biosecurity measures are in place to prevent predator introduction to new islands, and predation does not threaten long-term viability of the populations in Delisting Criterion 1 because: a) nonnative snail predators do not exist on the island or islet where the population occurs; b) effective predator control with long-term management commitment has successfully reduced predation pressure such that population viability is maintained; or c) quantitative data on demography and predator/prey dynamics in the population indicate that the population will maintain long-term viability without active predator control. At least 6 of the 12 populations in Delisting Criterion 1 must exist on islands or islets where surveys confirm Euglandina and Platydemus manokwari are absent.

Delisting Criteria Rationale: Similarly to the discussion under Downlisting Criteria above, monitoring of snail populations over 10 years would allow tracking of population status and trend over

several additional generations. After downlisting criteria are met, documenting an additionally increased number of populations, with representation in suitable habitat on multiple islands, would indicate a substantial shift toward improved conservation status and indicate the species is likely secure.

Because predation risk is the primary conservation concern for these species, delisting will require a clear understanding of nonnative predator distribution, abundance, and predator-prey dynamics with listed snail populations. At present, Eua zebrina appears to be at least persisting on Tutuila in the presence of Euglandina, and the viability of these populations needs to be carefully assessed. Snail populations that coexist with predators and have been clearly confirmed to be viable, with or without special management, could contribute to meeting delisting criteria. However, given the extensive history of partulid and achatinellid snail extirpations on Pacific islands (including sudden decline and loss of Achatinella populations that had previously persisted in the presence of predators), it remains prudent to maintain a diverse range of populations at sites where predation risk is completely absent. Thus, we expect that establishing and maintaining snail populations on predator-free islands or islets will continue to be a necessary part of recovering the species.

III. RECOVERY ACTIONS

This draft recovery plan identifies recovery actions needed to implement the recovery strategy and meet the recovery criteria. Implementation of a recovery action will depend on its priority, availability of funds and resources, coordination with partners, and complexity and logistical constraints. A broad action may have multiple components developed as needed to best coordinate recovery implementation. Specific project-level implementation of these actions will be accomplished through shorter-term activities. These activities will be described collectively in a companion document, referred to as the Recovery Implementation Strategy, developed in coordination with the recovery partners interested and willing to work on implementing the activities. Activities are intended to be adaptable and guide recovery partners to coordinate recovery implementation and further describe those responsible for each action described in the plan. Because the Recovery Implementation Strategy will be a flexible working document, the activities it describes may be modified as needed without requiring future revision of the recovery plan itself, so long as they are consistent with the recovery plan.

As discussed in the Introduction, this draft recovery plan is a guidance document rather than being regulatory in nature. As such, implementation of recovery actions is voluntary and depends on the cooperation and commitment of numerous partners in this conservation effort. Note that all Federal agencies have an obligation under section 7(a)(1) of the Act to carry out programs for conservation of federally listed species.

The actions needed to alleviate threats to the species and achieve recovery criteria are organized below into five categories: (1) Survey and protect management units, (2) Control species-specific threats; (3) Expand existing wild populations and establish additional populations; (4) Develop regulations and policy to support species recovery; and (5) Conduct additional research.

1. Survey and protect management units

1.1. Identify and survey extant populations for all species and the habitats in which they occur to assess current distribution, abundance, and habitat use. For all species, surveys are needed to assess current distribution, abundance, and habitat use. For peʻapeʻa vai, surveys are needed for the known extant populations on smaller islands in Fiji (e.g., Taveuni Island, the Lau Group). For maʻomaʻo, surveys are needed on Savaiʻi and ʻUpolu in Sāmoa. For tuʻaimeo, surveys and demographic studies are currently ongoing on Ofu and Olosega, and should be continued to assess population status. A survey comparable to the 1998 survey (Cowie and Cook 1999) is necessary to assess status of Eua zebrina (as well as other partulid snail species) across the American Sāmoa islands and determine whether there have been changes. To track ongoing status and trend in populations and identify any emerging issues, these surveys should be incorporated into an ongoing monitoring program, implemented cooperatively with partners and periodically repeated with consistent and comparable methodology.

1.2. Survey historically occupied areas for any persisting populations. In addition to surveys of extant populations, surveys are also needed in areas that they are known to have been occupied historically in order to determine whether populations persist. This includes, but is not limited to, surveys of areas within Sāmoa, American Sāmoa (particularly Taʻu), Vanuata, Tonga, and other islands in Fiji for the peʻapeʻa vai; and Tutuila for the maʻomaʻo; as well as assessing whether Ostodes strigatus is still extant on Tutuila. 1.3. Develop fine-scale climate models for these species to identify future suitable habitat based on existing and historical ranges and to determine potential future climate conditions. For all species, habitat assessment is needed to determine and accurately identify suitable habitat requirements for all life stages. This assessment will be used to evaluate the direct and indirect effects of various aspects of climate change (e.g., temperature, storm frequency and severity, sea-level rise, etc.) on the species. Fine-scale models can identify where important habitat needs to be protected to recover the species. Direct sensitivity to microclimate is likely to be a significant consideration for snail habitat and for peʻapeʻa vai roost sites; storm impacts can be significant for all species. 1.4. Identify and prioritize areas necessary for habitat protection and restoration. Within management units identified in this recovery plan, specific sites should be identified where habitat protection and restoration actions are the highest priority. Critical habitat was not formally designated for these species at the time of listing.

1.5. Ensure long-term protection of management units. Develop and implement actions necessary to ensure long-term protection of management units for all listed species. Specific actions are detailed below:

1.5.1. Identify threats specific to the management units. For each management unit, planning for protection should incorporate the best available information from current surveys, occurrence and distribution of listed species, assessment of habitat availability, and specific threats to species and their habitat.

1.5.2. Within management units, construct and maintain fencing and remove ungulates, and identify a long-term funding stream to keep fenced areas free of ungulates. Ungulates affecting habitat are primarily pigs, as well as goats in some locations. Coordinating with local partners, islands and localities where ungulate impacts to listed species habitat are significant should be identified; at sites where fencing or removal of ungulates is feasible this should be implemented in cooperation with partners.

1.5.3. Control or eradicate habitat-modifying invasive plants within management units. Priority control efforts for invasive plants should be focused on eradicating incipient populations of known high-risk invasive plants, and controlling established populations of invasive plants that modify-habitat. Species of particular concern include but are not limited to. Adenanthera pavonina (lopa) and Falcataria moluccana (tamaligi, albizia). 1.5.4. Protect management units from human disturbance as necessary. Human disturbance was identified as a threat primarily for peʻapeʻa vai roost sites. Protection of caves and surrounding vegetated areas should be implemented as necessary. This can be done by limiting public access through management restrictions and signage, discontinuing guano mining activities, gating of caves if feasible, providing outreach and education on human disturbance, or other means. 1.5.5. Protect forest habitat from disturbance, clearing, and fragmentation. Because all of the listed species use mature forest habitat, maintaining existing habitat in intact condition, restoring cleared or storm-damaged areas, and connecting forest patches to minimize fragmentation should be implemented through working cooperatively with land management agencies, village chiefs and councils, and private landowners.

1.5.6. Control other threats as appropriate. If new threats to management units are identified, take appropriate measures to control the threats. 1.6 Monitor management and use results to adapt management actions. Implement an appropriate monitoring strategy for all management units, including monitoring to assess threat status and species responses to management. Use an adaptive management strategy to adjust management based on observed outcomes.

2. Control species-specific threats.

2.1. Develop and implement control programs for cats and rats, updating methods as new technology becomes available. Work with stakeholders to develop and implement the most appropriate strategy for effective rodent and cat control programs within various management units, considering cost-effectiveness, conservation benefit for each listed species, and local support for actions. Predator control programs for cats and rats should consider use of all currently available tools and methods, such as trapping grids using snap traps or automatic

traps (e.g., Goodnature A24 automatic traps), rodenticide applications, lethal removal or relocation of cats, eradication plans on smaller islets with traps and/or rodenticide, or control measures at sources such as ports or agricultural locations. Research on developing and assessing the effectiveness of new control methods (e.g., birth control, gene drive technology, etc.) should also be explored as new technology becomes available to expand options for predator control tools.

2.2. Develop and implement control programs for nonnative invertebrates (Euglandina, Platydemus manokwari), updating methods as new technology becomes available. Currently, control methods for nonnative invertebrates consist of ground searches and hand removal, which is labor intensive and time costly. Predator-proof exclosures in Hawaiʻi have been shown to be effective (although not absolutely) in excluding Euglandina (and rats) from limited areas, but there is currently no known design that will also exclude Platydemus manokwari. Further research is needed to develop an exclosure design that will exclude both predators (in addition to rats). Furthermore, as new technology becomes available, other methods such as gene drive or other genetic approaches should be researched and assessed for their feasibility and effectiveness for landscape control of nonnative invertebrates.

2.3. Control other threats to specific species as appropriate. If new threats are identified, implement the appropriate measures to control the threats.

2.4. Monitor management and use results to adapt management actions. Implement an appropriate monitoring strategy for all species, including monitoring of threats to assess effectiveness of control actions. Use an adaptive management strategy to adjust management based on observed outcomes. 3. Expand the distribution of existing wild populations and establish additional populations. Planning and implementation of any translocation efforts should consider baseline information, review efforts for comparable species, and incorporate post-translocation monitoring (for overview of considerations, see IUCN 2013). 3.1. Identify areas within management units appropriate for establishing or augmenting populations, using information on climate and habitat suitability. For all species, fine-scale climate models (Recovery Action 1.2) can assist to identify the area(s) most suitable for establishing or augmenting populations within management units.

3.2. Conduct an assessment or feasibility study on translocation, husbandry, and captive propagation.

A thorough assessment or feasibility study should be conducted for the ability to implement translocation or captive propagation strategies for all species. In particular, strategies for the maʻomaʻo and peʻapeʻa vai are anticipated to be complex due to issues such as international political boundaries and limited species information. Another consideration for these strategies is the feasibility and conservation benefit of translocating species outside of their known historical range.

3.3. Select populations for augmentation or sites for reintroduction. Translocation efforts should focus on sites that are identified as high priority for augmentation or reintroduction.

3.4. Prepare reintroduction sites. Prior to translocation actions, reintroduction sites should be prepared so that they can reasonably sustain species populations. Site preparation measures should include, but are not limited to, implementing a predator control program, habitat restoration, and addressing other species specific threats.

3.5. Translocate individuals for augmentation or reintroduction with captive propagation if necessary. Selection of individuals to be translocated should consider their genetic characteristics and the potential effects of removal on the source population. 3.6. Release translocated individuals. Methods of releasing individuals (soft versus hard release, acclimation time, etc.) should be determined considering lessons learned from the outcome of any translocation efforts for similar species.

3.7. Monitor success of releases and use results to adapt management actions Any successes or failures noted in the release process or in subsequent monitoring should be considered when implementing subsequent steps for translocation projects.

4. Utilize regulations and policy to support species recovery 4.1. Support protection of endangered species under Territorial law Territorial law (American Sāmoa Code Annotated, Title 24) currently establishes a Natural Resources Commission to consider endangered species issues. If this legislation were amended to clarify its authority to carry out programs for the conservation of endangered species, the Territory would be eligible to establish a cooperative agreement under section 6 of the Act and receive funds for endangered species conservation projects through the section 6 grants program. This potential funding source could substantially contribute to accomplishing various recovery actions identified in this recovery plan.

36 4.2. Develop and implement a biosecurity plan to prevent the influx of new pests and invasive species into the Territory and between islands. Biosecurity measures are critical to avoid introducing snail predators to new islands within American Sāmoa, prevent reintroduction of rats if eradication programs are done, and intercept or control any additional invasive species that may be brought in from outside American Sāmoa. These might include invasive plants, brown tree snakes, or additional Euglandina clades. These measures should include outreach to members of the public who travel among islands, and should be developed and implemented with coordination among DMWR, airport and port managers, Coast Guard, and Army Reserve representatives. For detailed discussion of biosecurity implementation on Pacific islands, please see SPREP (2009). 5. Conduct additional research essential to recover species and restore the habitats on which they depend. Research results should be evaluated and used to adaptively modify implementation of recovery actions on the ground. 5.1. Assess current distribution of snail predators. A comprehensive assessment of the current distribution of introduced predators (rats, Platydemus manokwari, Euglandina, and other predatory snails) across the American Sāmoa islands and smaller islets is necessary to effectively plan recovery actions and any translocations for the listed snails.

5.2. Conduct studies on the habitat, demography, and dispersal of each species. Basic information valuable to inform recovery implementation would include: For peʻapeʻa vai, home range size, habitat and microclimate associations for roosts and foraging localities, and evaluation of pesticide or disease impacts; For maʻomaʻo, territory size, preferences for species composition and structure of forest habitat, and the quality and extent of available habitat at potential translocation sites; For tuʻaimeo: improved survey and population assessment methodologies; and For the snails, quantitative assessment of predator-prey interactions, mortality rates, microclimate requirements, and identification of any refugia. 5.3. Conduct a genetic study of all populations of Eua zebrina Known populations of Eua zebrina on Tutuila and Ofu should be assessed to determine whether management subunits can be identified that are genetically or morphologically distinctive. If such subunits exist, representation of this diversity should be considered in planning recovery implementation and translocation activities, and for evaluating whether recovery criteria are met.

5.4. Conduct population viability analyses (PVA) for each species as data becomes available. Demographic data on survivorship; reproductive success; dispersal, immigration and emigration; and annual variation in parameters may be used to develop a PVA, analyze sensitivity of parameters to uncertainty, and project probability of population persistence over time.

38 Table 5. Crosswalk relating threats, recovery criteria, and recovery actions.

Downlisting and Listing Factor Threat Recovery Actions Delisting Criteria

Peʻapeʻa Vai

Deforestation - A agricultural/urban 2 1.3, 1.4, 1.5.5, 3.4, 5.2 development Present or Threatened Destruction, Modification or Curtailment of its Goats 2 1.5.2, 3.4 Habitat or Range Sea-level rise 2 1.3, 1.4

B N/A

Overutilization

Predation by rats 3 2.1, 3.4 C Predation by cats 3 2.1, 3.4 Disease or Predation Disease (possible) 3 5.2

D Need for biosecurity planning and territorial 3 4.1, 4.2 Inadequacy of Existing Regulatory endangered species Mechanisms legislation Human disturbance of 2 1.5.4, 3.4 roosts Cyclone impacts 1.3, 3.1, 3.2, 3.3, 3.5, 3.6, 1 3.7 E Low population size 1.1, 1.2, 3.1, 3.2, 3.3, 3.5, 1 Other Natural or Manmade Factors 3.6, 3.7, 5.4

Breakdown of 1.1, 1.2, 3.1, 3.2, 3.3, 3.5, 1 metapopulation dynamics 3.6, 3.7, 5.4

Pesticides (possible) 3 5.2

Maʻomaʻo

Deforestation - A agricultural/urban 2 1.3, 1.4, 1.5.5, 3.4, 5.2 development Present or Threatened Destruction, Modification or Curtailment of its Invasive plants 2 1.5.3, 3.4, 5.2 Habitat or Range Pigs 2 1.5.2, 3.4, 5.2

B N/A

Overutilization

Downlisting and Listing Factor Threat Recovery Actions Delisting Criteria

C Predation by rats 3 2.1, 3.4

Disease or Predation Predation by cats 3 2.1, 3.4

D Need for biosecurity planning and territorial 3 4.1, 4.2 Inadequacy of Existing Regulatory endangered species Mechanisms legislation Cyclone impacts 1.3, 3.1, 3.2, 3.3, 3.5, 3.6, 1 E 3.7

Other Natural or Manmade Factors Low population size 1.1, 1.2, 3.1, 3.2, 3.3, 3.5, 1 3.6, 3.7, 5.4

Tuʻaimeo

A Deforestation - agricultural/urban Present or Threatened Destruction, development 2 1.3, 1.4, 1.5.5, 3.4 Modification or Curtailment of its Habitat or Range

B N/A

Overutilization

C Predation by rats 3 2.1, 3.4

Disease or Predation Predation by cats 3 2.1, 3.4

D Need for biosecurity planning and territorial 3 4.1, 4.2 Inadequacy of Existing Regulatory endangered species Mechanisms legislation Cyclone impacts 1.3, 3.1, 3.2, 3.3, 3.5, 3.6, 1 E 3.7

Other Natural or Manmade Factors Low population size 1.1, 1.2, 3.1, 3.2, 3.3, 3.5, 1 3.6, 3.7, 5.2, 5.4

Eua zebrina and Ostodes strigatus

B Collection 3 4.1 Overutilization

Downlisting and Listing Factor Threat Recovery Actions Delisting Criteria

Predation by rats 3 2.1, 3.4, 5.2 C

Disease or Predation Predation by nonnative 3 2.2, 3.4, 5.1, 5.2 invertebrates D Need for biosecurity planning and territorial 3 4.1, 4.2 Inadequacy of Existing Regulatory endangered species Mechanisms legislation Cyclone impacts 1.3, 3.1, 3.2, 3.3, 3.5, 3.6, 1 3.7 E Low population size 1.1, 1.2, 3.1, 3.2, 3.3, 3.5, Other Natural or Manmade Factors 1 3.6, 3.7, 5.4

Loss of local genetic 1 5.2 diversity

IV. TIME AND COST ESTIMATES

Achieving the recovery criteria for all five of these species is estimated to require, at minimum, 25 to 60 years.

Presented below is a table of site-specific recovery actions and their estimated costs of implementation. The cost table contains the estimated costs for each action, projected to the estimated date of delisting. Estimated costs include only project specific contract, staff, or operations costs in excess of base budgets. They do not include budgeted amounts that support ongoing agency staff responsibilities. This recovery plan does not commit the Service or any partners to carry out a particular recovery action or expend the estimated funds.

Estimated costs described in Table 6 incorporate planning, design, implementation, and research, monitoring, and evaluation associated with specific actions. Adaptive management actions evaluate the implementation of those actions to ensure that management/conservation tools are appropriately and effectively addressing impacts to the species and meeting the objective of this recovery plan. If the tools are not effective, changes in management should be made and additional planning and scientific research may be necessary.

42 Table 6. Priority and estimated cost of recovery actions.

Recovery Action Estimated Species Actions Priority Total Cost Addressed Notes 1. Survey and 1 $21M+ All Species surveys: ~$225K/yr protect Habitat protection & restoration management (reforestation primarily in Sāmoa and on units Tutuila, ungulate fencing and invasive plant control, bat roost protection in Fiji): ~$250K/yr, much of which is outside U.S. jurisdiction 2. Control 1 $106M+ All Rat and cat control (Ofu, Olosega, priority species- sites in Sāmoa, bat roost sites, possibly specific threats translocation localities): minimally ~$1,500K/yr assuming current methods, perhaps significantly greater, most of which is outside U.S. jurisdiction Invertebrate control (research on new control technologies, and/or localized removal): ~$1,000K/year 3. Expand the 1 $1.7 M All Translocation (Eua zebrina , maʻomaʻo, distribution of and tuʻaimeo are most likely; sites to be existing wild chosen and effort are TBD; includes populations planning, implementation, and and establish monitoring): ~$1,700K additional populations 4. Develop 2 $180M All Cooperative agreement (American regulations and Sāmoa): administrative action policy to Biosecurity implementation (American support species Sāmoa): ~$3,000K/yr recovery 5. Conduct 1 $1M All Snail predator survey (included in snail additional survey); demographic studies (Eua research zebrina); genetics analysis (Eua zebrina); essential to population viability analyses (all species): recover species ~$1,000K and restore habitats Estimated cost through date of recovery (25-60 years by species, as discussed below): $310M+

Cost estimates are preliminary. Project-level details of recovery action implementation are to be developed cooperatively with partners in a Recovery Implementation Strategy for this recovery plan. Implementation is subject to availability of funds and is at the discretion of partners and sovereign governments.

43 Priority 1 Action: An action that must be taken to prevent extinction or prevent the species from declining irreversibly in the foreseeable future Priority 2 Action: An action that must be taken to prevent a significant decline in species population or habitat quality. Priority 3 Action: All other actions necessary to meet the recovery objectives.

Date of Recovery: If all actions are fully funded and implemented as outlined, including full cooperation of all partners needed to achieve recovery, then we estimate the earliest that the delisting criterion could be met would be the year 2060 for peʻapeʻa vai, 2080 for maʻomaʻo, 2050 for tuʻaimeo, and 2045 for Eua zebrina.

For peʻapeʻa vai, delisting criteria include a 20-year monitoring period, and resettlement of any substantial portion of its historical range from the remnant population in Fiji is likely to require some additional decades of active and successful management.

Maʻomaʻo, which likely has a long generation time and will need extensive predator control and possibly a translocation program, has delisting criteria that include a 40-year monitoring period and will probably require an additional several decades of management for work across national boundaries to fully reverse declining population trends.

Tuʻaimeo, which probably has a shorter generation time but is difficult to survey accurately, will primarily require predator control to be implemented on Ofu and Olosega along with improved survey methods, perhaps in combination with translocation to other islands; if efforts are successful, delisting criteria could be achieved within 30 years.

Eua zebrina still appears to be widely distributed at least on Tutuila, but requires current survey data, better understanding of predator dynamics, and potentially translocation to other islands; if results of these are favorable, delisting might be achieved within 25 years.

Recovery of Ostodes strigatus cannot be reasonably projected unless a population is rediscovered, but in any scenario recovery is unlikely to be reached before the year 2060.

V. REFERENCES A. LITERATURE CITED

[ASCC] American Samoa Community College. 2010. American Samoa forest assessment and resource strategy 2011-2015. Prepared by Forestry Program, Division of Community and Natural Resources. 63 pages. Amerson, A.B., Jr., W.A. Whistler, and T.D. Schwaner. 1982. Wildlife and wildlife habitat of American Samoa. II. Accounts of flora and fauna. U.S. Fish and Wildlife Service, Washington, D.C. 151 pages. Atherton, J. and B. Jefferies. 2012. Rapid biodiversity assessment of upland Savaii, Samoa. Apia, Samoa. 176 pages. [ABOM and CSIRO] Australian Bureau of Meteorology and Commonwealth Scientific and Industrial Research Organization. 2011. Climate Change in the Pacific: Scientific assessment and new research. Volume 2: Country Reports. Chapter 12. Samoa. Baptisa, L.F., P.W. Trail, and H.M. Horblit. 1997. Family Columbidae (Pigeons and Doves). Pages 60-243 in del Hoyo, J., A. Elliott, and J. Sargatal editors. 1997. Handbook of the Birds of the World. Volume 4. Sandgrouse to Cuckoos. Lynx Edicions, Barcelona. BirdLife International. 2016. Gymnomyza samoensis. The IUCN Red List of Threatened Species 2016: e.T22704317A93962858. http://dx.doi.org/10.2305/IUCN.UK.2016- 3.RLTS.T22704317A93962858.en. Downloaded on 16 May 2019. Butler, D. and R. Stirnemann. 2013. Leading the recovery of two of Samoa’s most threatened bird species, the tooth-billed pigeon (Manumea) and the mao (Mao) through ecological research to identify current threats. Biodiversity Conservation Lessons Learned Technical Series 25. 50 pages. Clarke, B., J. Murray, and M.S. Johnson. 1984. The extinction of endemic species by a program of biological control. Pacific Science 38:97-104 Clarke, D., Editor. 2019. EAZA best practice guidelines for Polynesian tree snails (Partula spp.). Edition 1.0. Partula snail EEP Species Committee. June 2019. 61 pages. Clunie, F. 1984. Birds of the Fiji bush. Fiji Museum, Suva, Fiji. 158 pages. Cowie, R.H. 1992. Evolution and extinction of Partulidae, endemic Pacific island land snails. Philosophical Transactions: Biological Sciences 335:167-191. Cowie, R.H. 1998. Catalog of the nonmarine snails and slugs of the Samoan islands. Honolulu, Hawaii. Bishop Museum Press. 122 pages. Cowie, R.H. 2001. Decline and homogenization of Pacific faunas: the land snails of American Samoa. Biological Conservation 99:207-222.

45 Cowie, R. and R.H. Cook. 1999. The distribution and abundance of land snails in the National Park of American Samoa, with particular focus on Partulidae. Report. Cooperative National Parks Resources Studies Unit, University of Hawaii at Manoa. 142 pages. Cowie, R. and R.H. Cook. 2001. Extinction or survival: Partulid tree snails in American Samoa. Biodiversity and Conservation 10:143-159. Cowie, R.H., R.J. Rundell, F. Mika, and P. Setu. 2003. The endangered partulid tree snail Samoana thurstoni (Cooke and Crampton, 1930) on Olosega and the land snail diversity of the Manua Islands, American Samoa. American Malacological Bulletin 17:37-43. Craig, P., Editor. 2009. Natural history guide to American Samoa. 3rd Edition. National Park Service. 130 pages. [DMWR] Department of Marine and Wildlife Resources, American Samoa Government. 2006. A comprehensive strategy for wildlife conservation in American Samoa. September 2006, Revised November 2006. 116 pages. Eldredge, L.G. 1988. Case studies of the impacts of introduced animal species on renewable resources in the U.S.-affiliated Pacific Islands. In: B.D Smith, Editor. Topic reviews in insular resource development and management in the Pacific U.S.-affiliated islands. University of Guam Marine Lab. Engbring, J. and F.L. Ramsey. 1989. A 1986 survey of the forest birds of American Samoa. U.S. Fish and Wildlife Service, Department of Interior, Honolulu, Hawaii. Esselstyn, J.A., G.J. Wiles, and A. Amar. 2004. Habitat use of the Pacific sheath-tailed bat (Emballonura semicaudata) on Aguiguan, Mariana Islands. Acta Chiropterologica 6:303-308. Geering, D. 2002. Longevity of Aust birds. Accessed 30 May 2019 at http://bioacoustics.cse.unsw.edu.au/birding-aus/2002-04/msg00467.html Girardi, E.L. 1978. The Samoa land snail genus Ostodes (Mollusca: Prosobranchia: Poteriidae). The Veliger 20:191-250. Grant, G.S., S.A. Banack, and P. Trail. 1994. Decline of the sheath-tailed bat Emballonura semicaudata (Chiroptera: Emballonuridae) on American Samoa. Micronesica 27:133- 137. Hadfield, M.G., S.E. Miller, and A.H. Carwile. 1993. The decimation of endemic Hawaiian tree snails by alien predators. American Zoologist 33:610-622. Hansen, M.C., P.V. Potopav, R. Moore, M. Hancher, S.A. Turubanova, A. Tyukavina, D. Thau, S.V. Stehman, S.J. Goetz, T.R. Loveland, A. Kommareddy, A. Egorov, L. Chini, C.O. Justice, and J.R.G. Townshend. 2013. High-resolution global maps of 21st century forest cover change. Science 342:850-853. Data available: http://earthenginepartners.appspot.com/science-2013-global-forest

46 Helgen, K.M. and T.F. Flannery. 2002. Distribution of the endangered Pacific sheathtail bat Emballonura semicaudata. Australian Mammalogy 24:209-212. Higgins, P.J., L. Christidis, and H.A. Ford. 2008. Family Meliphagidae (Honeyeaters). Pages 498-691 in del Hoyo, J., A. Elliott, and D.A. Christie editors. 2008. Handbook of the Birds of the World. Volume 13. Penduline-tits to Shrikes. Lynx Edicions, Barcelona, Spain. Hutson, A.M., S.P. Mickleburgh, and P.A. Racey (compilers). 2001. Microchiropteran bats: global status survey and conservation action plan. IUCN/SSC Chiroptera Specialists Group, International Union for the Conservation of Nature and Natural Resources, Gland, Switzerland and Cambridge, United Kingdom. [IUCN] International Union for the Conservation of Nature. 2013. Guidelines for reintroductions and other conservation translocations Version 1.0. Gland, Switzerland. IUCN Species Survival Commission. viii +57 pages. Jackson, D.B. and R. Jit. 2007. Population densities and detectability of 3 species of Fijian forest birds. Notornis 54:99-111. Kayano, K., P. Pyle, and D. Kesler. 2019. Development of remote tracking and vocalization playback methodology to study the natural history of Tongan ground dove on Ofu and Olosega Islands, American Samoa. The Institute for Bird Populations, Point Reyes Station, California. 28 pages. Koopman, K.R. and D.W. Steadman. 1995. Extinction and biogeography of bats on ʻEua, Kingdom of Tonga. American Museum Novitates 3125:1-13. 1995 Malotaux, J. 2012. Bat caves in Fiji: Status and conservation of roosting caves of the Fiji blossom bat, the Pacific sheath-tailed bat, and the Fiji free-tailed bat. NatureFiji - MareqetiViti. 40 pages. McLaughlin, G.M. and P.K. Dearden. 2019. Invasive insects: management methods explored. Journal of Insect Science 19:1-9. Meyer, W.M. III, N.W. Yeung, J. Slapcinski, and K.A. Hayes. 2017a. Two for one: inadvertent introduction of Euglandina species during failed bio-control efforts in Hawaii. Biological Invasions 19:1399-1405. Meyer, R., J. Seamon, S. Faamu, and I. Lalogafuafua. 2017b. Classification and mapping of wildlife habitat in American Samoa: An object-based approach using high resolution orthoimagery and LIDAR remote sensing data. American Samoa Department of Marine and Wildlife Resources. 141 pages. Mickleburgh, S.P., A.M. Hutson, and P.A. Racey. A review of the global conservation status of bats. Oryx 36:18-34. Miller, S.E. 1993. Final report on surveys of the arboreal and terrestrial snail fauna of American Samoa. University of Hawaiʻi. Report to U.S. Fish and Wildlife Service, Honolulu, Hawaiʻi. 30 pages.

47 [MNRE] Samoa Ministry of Natural Resources and Environment. 2006. Recovery plan for the maʻomaʻo or mao (Gymnomyza samoensis). 39 pages. [NMFS and USFWS] National Marine Fisheries Service and U.S. Fish and Wildlife Service. 2018. Interim Recovery Planning Guidelines, Version 1.4. 122 pages. O’Shea, T.J. and E.W. Valdez. 2009. Assessment for Pacific sheath-tailed bats (Emballonura semicaudata rotensis) on Aguiguan, Commonwealth of the Northern Mariana Islands. Administrative Report to U.S. Fish and Wildlife Service, U.S. Geological Survey, Fort Collins Science Center, Fort Collins, Colorado. 118 pages Ohbayashi, T., I. Okochi, H. Sato, T. Ono, and S. Chiba. 2007. Rapid decline of endemic snails in the Ogasawara Islands, Western Pacific Ocean. Applied Entomology and Zoology. 42:479-485. Ollier, C.D., W.A. Whistler, and A.B. Amerson, Jr. 1979. O le Pupu Pue National Park, Samoa Vol. 1 Main Report. Environment Dept MSS. W. Samoa. Otis, D.L., J.H. Schulz, D. Miller, R.E. Mirarchi, and T.S. Baskett (2020). Mourning Dove (Zenaida macroura), version 1.0. In Birds of the World (A. F. Poole, Editor). Cornell Lab of Ornithology, Ithaca, New York.https://doi.org/10.2173/bow.moudov.01 [PCCSP] Pacific Climate Change Science Program. 2014. Current and future climate of Samoa. 8 pages. Pacifici, M., L. Santini, M. Di Marco, D. Baisero, L. Francucci, G. Grottolo Marasini, P. Visconti, and C. Rondinini. 2013. Generation length for mammals. Nature Conservation 5:87-94. Palmeirim, J.M., A. Champion, A. Naikatini, J. Niukula, M. Tuiwawa, M. Fisher, M. Yabaki- Gounder, S. Qalovaki, and T. Dunn. 2005. Distribution, status and conservation of the bats in the Fiji Islands. University of the South Pacific. 69 pages. Palmeirim, J.M., A. Champion, A. Naikatini, J. Niukula, M. Tuiwawa, M. Fisher, M. Yabaki- Gounder, S. Throsteinsdottir, S. Qalovaki, and T. Dunn. 2007. Distribution, status and conservation of the bats of the Fiji Islands. Oryx 41:509-519. Pratt, H.D., P.L. Bruner, and D.G. Berrett. 1987. A field guide to the birds of Hawaii and the tropical Pacific. Princeton University Press, Princeton, New Jersey. 409 pages + plates. Rohrer, J., V. Costello, J. Tanino, L. Bialic-Murphy, J. Sprague, and S. Joe. 2013. Development of tree snail protection enclosures: from design to implementation. Draft Technical Report. Pacific Cooperative Studies Unit, University of Hawaii at Manoa. 33 pages. Ruffell, J. and S. Parsons. 2009. Assessment of the short-term success of a translocation of lesser short-tailed bats Mystacina tuberculata. Endangered Species Research 8:33-39. Ruffell, J., J. Guilbert, and S. Parsons. 2009. Translocation of bats as a conservation strategy: previous attempts and potential problems. Endangered Species Research 8:25-31.

48 Scanlon, A.T., S. Petit, and G. Bottrof. 2014. The conservation status of bats in Fiji. Oryx 48:451-459. [SPREP] Secretariat of the Pacific Regional Environment Programme. 2009. Guidelines for invasive species management in the Pacific: a Pacific strategy for managing pests, weeds, and other invasive species. Compiled by Alan Tye. Apia, Samoa. 20 pages. Space, J.C. and T. Flynn. 2002. Report to the Government of Samoa on invasive plant species of environmental concern. U.S.D.A. Forest Service, Pacific Southwest Research Station, Institute of Pacific Islands Forestry, Honolulu, Hawaii. 80 pages. Steadman, D.W. and H.B. Freifeld. 1998. Distribution and relative abundance of landbirds in the Vavau group, Kingdom of Tonga. Condor 100:609-628 Stirnemann, R., M.A. Potter, D. Butler, and E. Minot. 2015. Slow life history traits in an endangered tropical island bird, the Maʻomaʻo. Bird Conservation International 26:366- 379. Tarburton, M.K. 2002. Demise of the Polynesian sheath-tailed bat Emballonura semicaudata in Samoa. Micronesica 34:105-108. [USFWS] U.S. Fish and Wildlife Service. 2015. Endangered and threatened wildlife and plants: Proposed endangered status for five species from American Samoa; Proposed rule. October 13, 2015. Federal Register 80:61568-61607. [USFWS] U.S. Fish and Wildlife Service. 2016. Endangered and threatened wildlife and plants; Determination of endangered status for five species from American Samoa. December 9, 2016. Federal Register 81:65466-65508. [USFWS] U.S. Fish and Wildlife Service. 2019a. Recovery outline for American Samoa species. Portland, Oregon. 21 pages. [USFWS] U.S. Fish and Wildlife Service. 2019b. Endangered and threatened wildlife and plants; Initiation of 5-year status reviews for 91 species in Oregon, Washington, Hawaii, and American Samoa. June 11, 2019. Federal Register 84:27152-27154. [USFWS] U.S. Fish and Wildlife Service. 2019c. Amendment to the recovery plan for the Oahu tree snails of the genus Achatinella. August 7, 2019. 11 pages. [USFWS] U.S. Fish and Wildlife Service. 2020a. Pacific sheath-tailed bat (Emballonura semicaudata semicaudata) species report. Pacific Islands Fish and Wildlife Office, Honolulu, Hawaiʻi. 36 pages. [USFWS] U.S. Fish and Wildlife Service. 2020b. Mao (Gymnomyza samoensis) species report. Pacific Islands Fish and Wildlife Office, Honolulu, Hawaiʻi. 32 pages. [USFWS] U.S. Fish and Wildlife Service. 2020c. Friendly ground-dove (Alopecoenas [Gallicolumba] stairi, tuaimeo) species report. Pacific Islands Fish and Wildlife Office, Honolulu, Hawaiʻi. 40 pages.

49 [USFWS] U.S. Fish and Wildlife Service. 2020d. Eua zebrina species report. Pacific Islands Fish and Wildlife Office, Honolulu, Hawaiʻi. 33 pages. [USFWS] U.S. Fish and Wildlife Service. 2020e. Ostodes strigatus species report. Pacific Islands Fish and Wildlife Office, Honolulu, Hawaiʻi. 34 pages. Watling, D. 2001. A guide to the birds of Fiji and Western Polynesia including American Samoa, Niue, Samoa, Tokelau, Tonga, Tuvalu and Wallis and Futuna. Environmental Consultants, Suva, Fiji. Wiles, G.J., T.J. O’Shea, D.J. Worthington, J.A. Esselstyn, and E.W. Valdez. 2011 Status and natural history of Emballonura semicaudata rotensis on Aguiguan, Mariana Islands. Acta Chiropterologica 13:299-309. Wiles, G.J. and D.J. Worthington. 2002. A population assessment of Pacific sheath-tailed bats (Emballonura semicaudata) on Aguiguan, Mariana Islands. Prepared for the U.S. Fish and Wildlife Service, Pacific Islands Fish and Wildlife Office, Honolulu, Hawaii. 32 pages. Wiley, J.W. and J.M. Wunderle, Jr. 1993. The effects of hurricanes on birds, with special reference to Caribbean islands. Bird Conservation International 3:319-349.

B. IN LITT. REFERENCES

Miles, Adam. 2015. Sheath-tailed bat or other species data. Email from Adam Miles, Department of Marine and Wildlife Resources, to James Kwon, U.S. Fish and Wildlife Service. February 11, 2015. Thomas-Moko, Nunia. 2020. Sheath-tailed bat recovery planning. Email from Nunia Thomas- Moko, NatureFiji, to Grant Canterbury, U.S. Fish and Wildlife Service. March 10, 2020. Tye, Alan. 2012. CEPF final project completion report: Restoration of Nuutele and Nuulua islands. Secretariat of the Pacific Regional Environment Programme. 33 pages.