VU
Micronesian Imperial-pigeon Ducula oceanica



Justification

Justification of Red List Category
Although still reasonably abundant on some islands in its range, the species has undergone rapid declines on some islands, most notably on Palau and Pohnpei, as a result of hunting and habitat loss. Although data indicates a recent recovery on Pohnpei, the potential start of a recovery on Palau and the recovery of at least one small subpopulation, and overall trends are uncertain, a population reduction of over 30% is likely to have taken place over the last three generations. For this reason, the species is listed as Vulnerable.

Population justification
In Palau, the population in 2014 was estimated to be 3,000 individuals (Beouch 2016 in NEPC 2017). Since then, the trend has been unclear; experts considered that the population was declining (Eberdong and Kitalong pers. comm. in NEPC 2019), but data on eBird suggested a possible recovery from 2015. Eight-minute count data from the 20 monitoring stations showed a fall in the number of birds per station from 3.65 in 2014 to 1.88 in 2016, followed by a rise to 2.8 in 2017 (NEPC 2019). Based on the 2014 population estimate, the range of trends since 2014 shown in different datasets (NEPC 2019) and extrapolating trends, the population size in 2020 is placed in the band 1,300 - 3,300 individuals.

In Pohnpei, variable circular plot surveys in 1983 produced a population size estimate of 822 individuals (Engbring et al. 1990). The survey took place during a severe drought, which may have enhanced detectability of the species (Buden 2000). Repeat surveys took place in 1994 and estimated a decline in relative abundance of approximately 73% since 1983 (Buden 2000). If this decline is applied to the 1983 population estimate, then the estimated population on Pohnpei in 1994 is c. 222 individuals. Surveys in 2012 resulted in a population estimate of 5,025 individuals based on habitat models linked to density functions (Oleiro and Kesler 2015). A density model produced population density estimates that ranged from 11.8 individuals per km2 in mangroves to 33 individuals per km2 at elevations of 400-600m, with an average density of 17 individuals per km(Oleiro 2014). Oleiro (2014) also re-analysed detection rates from 1983 and 1994, together with new data from surveys using a similar method in 2012, and reported an increase of 220% from 1994 to 2012 (Oleiro 2014, Oleiro and Kessler 2015). Applying this trend to the 2012 estimate and extrapolating forwards would produce a 2020 estimate of 8,427 individuals, assuming exponential growth. However, the report by Oleiro (2014) also presents relative detection rates by elevation band for each survey (1984, 1994 and 2012). If the relative detection rates are applied to the estimated 2012 population densities and scaled by the relative areas of the island that fall within each elevation band, the population in 1994 can be calculated as 2,746 individuals. This suggests that the population size has undergone an 83% increase from 1994 to 2012.  If this rate of increase is applied to the previous population estimates (Engbring et al. 1990, Buden 2000), and trends are extrapolated forward to 2020, the population on Pohnpei in 2020 can be estimated at 531 individuals. The population size on Pohnpei in 2020 is therefore highly uncertain and is here placed in the band 500 - 8,500 individuals.

On Kosrae, Engbring et al (1990) estimated a mean population density of 60 individuals per km2 and a population of 7,474 individuals in 1983. Hayes et al. (2016) reported hearing only a few individuals during several long hikes in the mountains of the island in 2013 and 2014. The population is therefore suspected to have declined, but less severely than the decline observed on Pohnpei. It is here placed in the band 1,700 - 5,400 individuals (assuming that it has declined at a constant rate of 15-50% over three generations).

On Chuuk, a survey in 1983 estimated 51 individuals (Engbring et al. 1990), but this may have been an underestimate (O'Brien in litt. 2017). If this figure is scaled up by the same magnitude as that of the Pohnpei estimate according to the population estimate and relative detection rates in Oleiro (2014), the population in 1983 may be estimated to be 265 individuals. The population on Chuuk is not thought to have increased greatly since 1983 (G. Wiles in litt. 2020). Assuming the population on Chuuk is stable, the population on 2020 is placed in the band 50-300 individuals.

A small population of c.25-30 individuals has been reported on Ant Atoll (Buden 1996a). Engbring et al. (1990) reported a density of 27 individuals per km2 on Kikalap Aru in 1983. Surveys in 1994-5 reported a density of 14.9 individuals per km2 (Buden 1996a). Extrapolating this trend forward would produce a population estimate of 6-7 individuals. The population on Ant Atoll is here placed in the band 6-20 individuals.

There were thought to be c.50 individuals on Sapwuahfik Atoll (Buden 1999) and small numbers have been reported on Pakin Atoll (70 individuals per kmin 1994, which may equate to 161 individuals; Buden 1996b) and on Satawan Atoll (Buden 2006). It is not known whether these populations persist. The total 2020 population estimate for these atolls is placed in the band 0-300 individuals.

The population on Nauru was roughly estimated at 75-100 individuals in 2006-2007 (Buden 2008). The species is thought to have declined on the island, and local people reported that they didn't see the species as often as they did previously (Buden 2008). It is now thought to be restricted to the Pinnacles area in the uplands, with more recent surveys suggesting a population of 50-150 individuals (Stirnemann 2015). Extrapolating these trends, the population in 2020 is placed in the band 30-240 individuals.

On Majuro Atoll in the Marshall Islands, a recovery programme led to an increase in numbers to c.80 individuals across 27 islets with a combined area of 5.2km2 (M. O'Brien in litt. 2011). The 2020 population is placed in the band 80 - 150 individuals. The populations on Lae, Lib and Namdrik, if still extant, must be small, as must the population on Tarawa; each is assumed to have a populations of no more than 50 individuals.

On Yap, Engbring et al. (1990) recorded a density of 23 individuals per km2  in 1983-1984 and a population of 572 individuals. Surveys in 2004 recorded no individuals along transects, but six individuals during inventories of the island, suggesting a possible decline since 1983 (Robinson and Robinson 2016). Since the population on Yap is considered to be introduced (Engbring et al. 1990), the population is not included in this assessment.

Overall, the population size is suspected to fall in the band 3,800 - 18,200 individuals, roughly equating to 2,500 - 12,200 mature individuals, with the best estimate placed at 8,000 mature individuals.

Although birds are known to fly among islands within island groups (Baptista et al. 1997, R. Davis in litt. 2020), there are likely to be at least seven subpopulations. Based on the population estimates above, the largest subpopulation may have between 1,700 - 8,600 individuals (assuming Pohnpei, Ant Atoll and Pakin Atoll may represent one subpopulation), roughly equating to 1,100-5,800 mature individuals.

Trend justification
In Palau, a National Bird Survey was carried out in 1991, following the US Fish and Wildlife Service's VCP (Variable Circular Plot) method and using DISTANCE analysis. This produced a population estimate of 13,718 individuals (Engbring 1992). Surveys were repeated in 2005 (VanderWerf 2007), but the resulting data were not analysed in the same way as those from the 1991 survey, and so are not comparable. However, relative abundance (birds/station) was compared between the two surveys, showing that abundance decreased by c.40% (from 1.46 to 0.87) between 1991 and 2005 (VanderWerf 2005). Declines were found on Babeldeob and in the Rock Islands. From 2005-2010, the National Bird Monitoring programme recorded a marked decrease in the number of monitoring stations at which individuals were observed (Olsen and Eberdong 2011), from 20 to 13 stations (Olsen and Eberdong 2011), suggesting a possible rapid decline in the population and/ or a possible restriction in range (Ketebengang and Gupta 2011), although most of these monitoring stations were on Babeldoab, so they may not provide a full picture of the national trend (R. Davis in litt. 2020). In 2014, the population on Palau was estimated to be 3,000 individuals (Beouch 2016 in NEPC 2017). In 2019, the trend was unclear; experts considered that the population was still declining (Eberdong and Kitalong pers. comm. in NEPC 2019), but data in eBird suggested a possible recovery from 2015 (NEPC 2019). Eight-minute count data from the 20 monitoring stations showed a fall in the number of birds per station from 3.65 in 2014 to 1.88 in 2016, followed by a rise to 2.8 in 2017 (NEPC 2019). Based on the 2014 population estimate and the range of available data for 2014-2018, the population size in 2016 is estimated to be between and 2,143 and 3,265 individuals, with 2,143 representing a best estimate. Based on the above information, and extrapolating to 2020 with assumed exponential decline, the population on Palau is estimated to have undergone a reduction of 75-86% over the past three generations (18 years) to 2020, with a best estimate of 86%.

In Pohnpei, the population size was estimated at 822 individuals in 1983 (Engbring et al. 1990). The survey took place during a severe drought, which may have enhanced detectability of the species (Buden 2000). Repeat surveys took place in 1994 and a decline in relative abundance was detected, estimated at approximately 73% since 1983 (Buden 2000). If this decline is applied to the 1983 population estimate, then the estimated population on Pohnpei in 1994 is c. 222 individuals. Surveys in 2012 resulted in a population estimated of 5,025 individuals based on habitat models linked to density functions (Oleiro and Kesler 2015). Oleiro (2014) re-analysed detection rates from 1994 and reported an increase of 220% from 1994 to 2012. However, the report by Oleiro (2014) also presents relative detection rates by elevation band for each survey (1984, 1994 and 2012). If the relative detection rates are applied to the reported 2012 population densities and scaled by the relative areas of the island that fall within each elevation band, the population in 1994 can be calculated as 2,746 individuals. This suggests that the population size increased by 83% from 1994 to 2012. Based on the rate of increase from 1994-2012 reported by Oleiro (2014), and extrapolating to 2020 with assumed exponential change, the population on Pohnpei can be estimated to have increased by 226% over the past three generations to 2020. Conversely, based on the rate of change since 1994 derived from Oleiro's (2014) population densities and relative detection rates, here considered the best estimate, the population is estimated to have increased by 85% over the past three generations.

On Kosrae, Engbring et al. (1990) estimated a population of 7,474 individuals in 1983. Hayes et al. (2016) reported hearing only a few individuals during several long hikes in the mountains of the island in 2013 and 2014. The population is therefore suspected to have declined, but less severely than the decline observed on Palau. In the last three generations to 2020, the population on Kosrae is suspected to have declined by 15-50%, with a best estimate of 30%.

On Chuuk, a survey in 1983 estimated 51 individuals (Engbring et al. 1990), but this may have been an underestimate (O'Brien in litt. 2017). This population is unlikely to have increased greatly since 1983 (G. Wiles in litt. 2020). Over the past three generation to 2020, the population on Chuuk is suspected to have declined by 0-50%, with a best estimate of no change.

A small population of c.25-30 individuals has been reported on Ant Atoll (Buden 1996a). Engbring et al. (1990) reported a density of 27 individuals per km2 on Kikalap Aru in 1983. Surveys in 1994-5 reported a density of 14.9 individuals per km2 (Buden 1996a). Extrapolating this rate of change would produce an estimated reduction of 63% over the past three generations. The reduction over the past three generations to 2020 is suspected to be between 25% and 63%, with a best estimate of 25%.

There were thought to be c50 individuals on Sapwuahfik Atoll (Buden 1999) and small numbers have been reported on Pakin Atoll (70 individuals per kmin 1994; Buden 1996b) and on Satawan Atoll (Buden 2006). It is not known whether the populations on these atolls are still extant, and the changes over the past three generations to 2020 are not known, but these populations are suspected to have declined by up to 50%, with a best estimate assumed to be no change.

The population on Nauru was roughly estimated at 75-100 individuals in 2006-2007 (Buden 2008). The species is thought to have declined on the island, and local people reported that they didn't see the species as often as they did previously (Buden 2008). It is now thought to be restricted to the Pinnacles area in the uplands, with more recent surveys suggesting a population of 50-150 individuals (Stirnemann 2015). Based on the figures from 2006-2007 and 2015 and extrapolating to 2020, the population change on Nauru over the last three generations could range from an 80% reduction to a 388% increase, with a best estimate placed at a 33% increase.

On Majuro Atoll in the Marshall Islands, a recovery programme has led to an increase in numbers to c.80 individuals across 27 islets with a combined area of 5.2 km2 (M. O'Brien in litt. 2011). The rate of increase is not known. Over the past three generations to 2020, the population is suspected to have increased by between 0 and 242% (based on an increase of 50-80 from 2004 to 2011), with a best estimate of a 51% increase. 

The population on Tarawa is assumed to have undergone a population change ranging from a 50% reduction to a 50% increase over the past three generations, with a best estimate of no change. The populations on Lae, Lib and Namdrik are assumed to have undergone a reduction of 0-100% over the past three generations.

Combining the pessimistic estimates (above) of the population change over the past three generations (18 years) to 2020, and weighting by population size, gives a reduction of 73%. Combining the optimistic estimates of the population change over the past three generations, and weighting by population size, above gives a reduction of 10%. Combining the best estimates of the population change over the past three generations, and weighting by population size, gives a best estimate of a 36% reduction. The population is suspected to continue to decline for at least several years in the future.

Although data suggest that the population size may start to increase over the next decades if current trends continue, this is highly uncertain, so the species is inferred to be undergoing a continuing decline in population size.

Distribution and population

This species occurs in the Micronesian islands of PalauFederated States of Micronesia, the Marshall Islands, Nauru (Buden 2008, Stirnemann 2013) and Kiribati (Tarawa; M. O'Brien in litt 2017, eBird 2020), including many offshore islands. In Palau, it occurs across much of the island group, including across Babeldaob, the Rock Islands and Sonsorol, where it was reportedly introduced (G. McKinlay in litt. 2020). In Federated States of Micronesia, it occurs on Pohnpei, Chuuk, Ant Atoll (Engbring et al. 1990, Buden 1996a, eBird 2020) and Kosrae, as well as Yap, where it is considered to have been introduced (Engbring et al. 1990). Small populations have also been reported in recent decades on Pakin and Sapwuahfik atolls in Pohnpei State, and Satawan atoll in Chuuk State (Buden 1996a, b, 1999, 2006). In the Marshall Islands, it is known to occur on Majuro (eBird 2020, G. Wiles in litt 2020), and has been reported to occur on Lae, Lib and Namdrik (National Biodiversity Team 2000 in Stennemann 2006). In Kiribati there are recent records on Tarawa (M. O'Brien in litt 2017, eBird 2020). It previously occurred on Kapingamarangi Atoll in Federated States of Micronesia (Buden 1998), Kuria and Aranuka in Kiribati (VanderWerf 2007), and across the Marshall Islands from Wotje to Jaluit and Mili (Vander Velde 2001, Baptista et al. 2020), but is now probably extinct from some of its former range.

Ecology

It is a forest species, occupying native lowland and montane forests, secondary forest, beach forest and mangroves and occasionally agroforest, although thickets of short hibiscus trees are avoided (Engbring et al. 1990, Engbring 1992, VanderWerf 2007, Oleiro and Kesler 2015). Surveys in Pohnpei indicated an association with old growth forest with higher canopies, and it was less common in disturbed habitat (Oleiro and Kesler 2015). Apparent patterns in habitat use are often more reflective of hunting pressure than habitat preference or quality (Engbring et al. 1990). It feeds on a variety of fruits and large fleshy seeds (Baptista et al. 1997). It is thought to breed year-round (Engbring et al. 1990, Baptista et al. 1997).

Threats

Hunting has long been the main threat to this species on nearly all of the islands in its range, with some subpopulations being overhunted (and extirpated from several atolls) by the early to mid-1900s (Baker 1951, Engbring et al. 1990, Buden 2006, Ketebengang and Gupta 2011, Hayes et al. 2016). Current hunting levels remain a major concern and are likely driving the declines noted on most islands (Engbring et al. 1990; Buden 1996b, 1999, 2008; Ketebengang and Gupta 2011; Hayes et al. 2016). Most hunting is likely done for subsistence use, but birds are also sold for profit, with nestlings occasionally taken for raising in captivity (VanderWerf 2007, Ketebengang and Gupta 2011, Hayes et al. 2016). Although many subpopulations receive various levels of legal protection, enforcement is weak and poaching is widespread (Engbring et al. 1990, Ketebengang and Gupta 2011, Hayes et al. 2016). Hunting has continued on Palau, including within protected areas, despite a fire-arms ban, and despite hunting the species being illegal (Engbring 1990, Buden 2000, Ketebengang and Gupta 2011, NEPC 2019, G. McKinlay in litt. 2020). The species continues to be hunted heavily on Kosrae, despite it being protected (Hayes et al. 2016, Department of Resources & Development, FSM 2019), and hunting has continued on Nauru, despite a firearms ban (Buden 2008). Hunting is also evident on Pohnpei (R. Davis in litt. 2020, G. McKinlay in litt. 2020).

Past and ongoing forest clearance for agriculture and development is another concern, particularly on Pohnpei and Chuuk (Buden 2000, Oleiro and Kesler 2015). In the Marshall Islands, significant areas of forest were cleared in the first half of the 20th century for military bases (M. N. Trevor in litt 2004). Upland forest on Pohnpei declined by about two thirds between 1975-1995, largely attributed to cultivation of sakau (also known as kava; Piper methysticum; Trustrum 1996, Merlin & Raynor 2005, Department of Resources & Development, FSM 2019). Forest is also cleared for road-building (Department of Resources & Development, FSM 2019). Most of the island's native forest has been either cleared or degraded to mixed forest with lowland secondary species (B. Raynor in litt. 2012). The clearance of areas of forest for sakau also introduces and facilitates the spread of invasive species (B. Raynor in litt. 2012, Department of Resources & Development, FSM 2019). Although efforts over the past 25 years to reduce the amount of clear-cutting for sakau plantations have resulted in the slowing of native forest conversion rates, the trend remains negative (Merlin and Raynor 2005, Oleiro and Kesler 2015). On Chuuk and Kosrae, forest has been lost to agriculture and the development of roads and housing, and wood is harvested for use as construction material and fuel (VanderWerf 2007, Department of Resources & Development, FSM 2019). On Nauru, forest has been lost due to phosphate mining, but restoration has been planned (Buden 2008). A study found that the species was negatively affected by reductions in forest canopy height due to agrcultural incursions (Oleiro and Kesler 2015).

Climate change is likely to cause sea level rise and increased storm frequencies and intensities in the region. An analysis of the affects of threats on D. oceanica populations found that sea level rise was likely to cause slow declines (Oleiro and Kesler 2015). Sea level rise is expected to inundate mangrove forests and could exacerbate forest loss in the uplands by forcing activities such as wood collection, agriculture and development to be moved to higher elevantions (Oleiro and Kesler 2015). Typhoons can also contribute to population declines. Pohnpei is currently relatively unaffected by typhoons, but could become more affected in the future. Climate change could also bring a greater risk of periodic drought to some islands, which could be especially harmful to subpopulations inhabiting atolls. For example, Buden (1998) noted that a former subpopulation on Kapingamarangi Atoll in Pohnpei State was eliminated during a severe drought in 1916-1918.

Conservation actions

Conservation and Research Actions Underway
The species is legally protected on Palau, Kosrae and the Marshall Islands. It occurs in a number of protected areas, including some traditional protected areas.
It is monitored on Palau as part of the National Program for Monitoring Forest and Coastal Birds. Campaigns against hunting have taken place and some enforcement of species protection laws is taking place (R. Davis in litt. 2020).
On Pohnpei, the forested upland interior is protected by the Pohnpei Watershed Forest Reserve, which was established in 1987. The law initially failed, but after involving communities via the Watershed Steering Committee in 1990, the Pohnpei Watershed Management Strategy was developed to ensure sustainable resource management (Merlin and Raynor 2005). The law gained support following public education and the reserve was delineated, to protect habitat from agricultural clearance, development and hunting. A campaign was conducted to encourage people to grow sakau at lower elevations, to protect the upland forests. Forest clearance has been monitored to evaluate success. Over 250 people have been appointed as voluntary 'community conservation officers' to help with conservation education and monitoring (Merlin and Raynor 2005).
In the Marshall Islands, a recovery programme for the species has been led by the Marshall Islands Conservation Society. Actions included reducing hunting and predation and increasing the number of fruiting trees (M. O'Brien in litt. 2011).

Conservation and Research Actions Proposed
Conduct surveys to assess its total population size and distribution. Monitor population trends through regular surveys, particularly in Chuuk and Kosrae (M. O'Brien in litt. 2017). Monitor rates of habitat loss and degradation. Monitor levels of hunting pressure.
Protect significant habitat, including by controlling development, and by creating and enforcing protected areas. Enforce laws against hunting, fire-arms and sales of protected species. Work with communities and island governments to manage the species sustainably, with outreach to island residents about the need for pigeon conservation being a priority (G. Wiles in litt. 2020).

Acknowledgements

Text account compilers
Wheatley, H.

Contributors
Ashpole, J, Buden, D., Davis, R.A., Derhé, M., Dutson, G., Gupta, A., Hayes, F.E., Mahood, S., Matthews, E., McKinlay, G., O'Brien, A., O'Brien, M., VanderWerf, E. & Wiles, G.


Recommended citation
BirdLife International (2022) Species factsheet: Ducula oceanica. Downloaded from http://www.birdlife.org on 11/08/2022. Recommended citation for factsheets for more than one species: BirdLife International (2022) IUCN Red List for birds. Downloaded from http://www.birdlife.org on 11/08/2022.