Endangered Species Recovery Efforts

Background

Reintroduction, the intentional movement of an organism into a part of its native range following its extirpation, is a common strategy in the conservation of threatened and endangered species (IUCN, 2013; Soorae, 2018). However, the success of this strategy in the past has been variable (Griffith, et al.

, 1989; Fischer & Lindenmayer, 2000; Wolf, et al., 1996), leading to calls for greater understanding of target species’ establishment and persistence requirements (Armstrong & Seddon, 2008; Seddon, et al., 2007). Habitat conditions are one of the largest contributors to reintroduction program outcomes (Griffith, et al., 1989). When animals depend on habitat patches that exhibit non-uniform spatiotemporal distribution of resources and conditions (Addicott, et al., 1987), they must move between patches to maximize their survival and fitness (Fretwell & Lucas, 1969; Morris, 2006). Furthermore, some species utilize multiple habitat types for different purposes or during different life stages (Wiens, 1976). Reintroduction programs that focus on such species can improve their likelihood of success by understanding the spatial and behavioral dynamics that occur within these habitats and how these dynamics change throughout the life stages of the species. Such knowledge will allow reintroduction practitioners to meet all the species’ ecological and behavioral needs and maximize the likelihood of persistence within the selected habitat.

Another challenge for reintroduction programs is the source of animals being released into the wild. Many reintroduction programs entail the release of captive-bred animals (Seddon, et al., 2007; Wolf, et al., 1996; Griffith, et al., 1989). However, these animals often exhibit behavioral differences when compared to wild-born animals, leading them to interact with their environment in different ways (Snyder, et al., 1996). These behavioral differences can be expressed in many contexts including spatiotemporal dynamics, resource exploitation skills, social interactions, and ultimately, survival (Mueller, et al., 2013; Snyder, et al., 1994; Meretsky, et al., 2000; Stoinski, et al., 2003). Understanding how released captive-bred animals, move through and utilize their habitat could be key to enhancing reintroduction program success.

Problem

The conservation program for the critically endangered Puerto Rican Parrot (Amazona vittata) incorporates both captive-breeding and reintroduction strategies (USFWS, 2009). In 1973, population managers began breeding parrots in captivity for release into the only extant wild population site, the El Yunque National Forest (hereafter El Yunque) (Snyder, et al., 1987). By 2006, managers began reintroducing a second wild population within the Rio Abajo State Forest (hereafter Rio Abajo), from where the species had been extirpated since the 1930’s (Snyder, et al., 1987; USFWS, 2009; PRDNER, 2016). Reintroduced birds were originally sourced from the captive-breeding program. However, the population is currently comprised of both captive-bred released birds and their wild-born descendants (PRDNER, 2016).

Despite the long history of the Puerto Rican Parrot Recovery Program (hereafter Recovery Program), only a handful of studies have been conducted on the species’ spatiotemporal dynamics. Furthermore, all the existing research on this topic has focused on the relic population within El Yunque (Snyder, et al., 1987; Lindsey, et al., 1991). Although early research included extensive habitat and diet characterization (Lugo & Frangi, 1993; Wunderle, 1999; Snyder, et al., 1987), El Yunque habitat differs greatly from existing and planned reintroduction sites on other parts of the island (Rivera & Aide, 1998; Ewel & Whitmore, 1973). Studies on other aspects of the species’ biology have been similarly restrictive in scope and location. The present study is the first to examine the spatiotemporal, social, and habitat use dynamics of the reintroduced population in Rio Abajo.

Study justification

Formerly abundant and widely distributed throughout Puerto Rico, the Puerto Rican Parrot (Amazona vittata; hereafter, PRP) is now considered one of the ten most endangered birds in the world (Raffaele et al. 1998). It is the only endemic psittacine species not yet extinct within the United States and its territories. The historical range of the PRP included the entire island of Puerto Rico and 3 satellite islands: Vieques, Culebra, and Mona (Snyder et al. 1987). The parrots on the islands of Culebra were recognized as distinct subspecies (Amazona vittata gracilipes). However, by 1940 a parrot population of approximately 2,000 birds (Rodriguez-Vidal 1959) was restricted to the Caribbean National Forest (CNF) located in eastern Puerto Rico. The population continued declining in subsequent years and approximately 14 parrots were thought to survive in 1975. However, with intense management, the wild flock increased at an average rate of 1.6 parrots per year from 1975 to 1987 (Lindsey et al. 1991). Currently, because of the impact of hurricanes, Irma and María there exist only one wild population of approximately 80 individuals within the karst region. Two additional captive populations are currently held in separate aviaries located in CNF and the Río Abajo Commonwealth Forest (RAF) in northcentral Puerto Rico. The aviary José L. Vivaldi located in RAF has been administered by the Commonwealth of Puerto Rico throughout the Puerto Rico Department of Natural and Environmental Resources (DNER) since 1992. These two facilities incorporate accepted methodologies for the management and safeguarding of the species in captivity and provide a sustainable source of parrots for fostering and future releasing of captive-reared parrots into the wild and to serve as genetic reserves in the advent of a natural catastrophe or disease outbreak.

The PRP is considered a national symbol with which most Puerto Ricans identify. The recovery programs of this species have had the intention to conserve and protect, not only the parrots but the natural resources associated with them. For example, the management practices implemented to ensure successful reintroduction of this species will also benefit the entire avian community and their habitat within this region. In the CNF, for example, the presence of this species has helped to maintain the protection that per years has had this natural reserve. Furthermore, the reintroduction of PRPs in the karst region helped protect this important region. This region contains the largest reported number of tree species per unit area in Puerto Rico and many of the endemic and endangered species of the island (Lugo et al. 2001). Also, this region contains the extensive freshwater aquifer of the entire island that serves as the main potable water supply for the population of Puerto Rico. For such reason, the recovery of this species within this area results in a benefit, as much for its habitat as for all the inhabitants of Puerto Rico.

During the last 40 years, conservation efforts have been underway to ensure the recovery of the species. These multifaceted recovery efforts include research, management, and captive breeding program (Wiley 1980, Collazo et al. 2010, White et al. 2014). To date, captive breeding is the primary method used to augment the PRP population (Collazo et al. 2010, White et al. 2000, White et al. 2014). However, as with most captive breeding programs, only a limited number of captive-produced parrots can be introduced annually into the wild population. As a result, there is a constant accumulation of parrots in captivity. Prolonged exposure to captivity can enhance the development of negative behavioral and physical effects (Derrickson and Snyder 1992, Wiley et al. 1992). Furthermore, hurricanes (e.g. Hugo-1989, Georges-1998, Irma-2017 and María-2017) have reduced the number of parrots in the wild (Vilella and García 1995). Wilson et al. (1994) reported that the wild population declined by 40% after hurricane Hugo. Also, hurricanes caused significant loss of available sites to cavity-nesting (Engstrom and Evans 1990). As a result of the apparent wild population’s vulnerability to extinction from natural and climatic threats, establishing an additional wild parrot population is critical to ensure a successful recovery of this endangered species (Lacey et al. 1989, Muiznieks 2001, FWS 1999, Llerandi-Román 2006, Collazo et al. 2010, Collazo et al. 2013, White et al. 2014).

Releasing captive-reared individuals to supplement the existing population or to re-establish other wild population has become a key component of the PRP recovery program and many endangered species recovery efforts (e.g., Biggins et al. 1998, Maloney and Murray 2000, Wanless et al. 2002, Brightsmith et al. 2005, Collazo et al. 2013, White et al. 2014). This recovery action minimizes the risks of the species’ extinction because it is less likely that catastrophic events (e.g., hurricanes), and other threats (e.g., predation), will affect equally and simultaneously spatially segregated populations. It is also hoped that releasing birds in a different environment will result in more vigorous population growth than what has characterized the extant wild population at the Caribbean National Forest over the past 40 years (e.g., 1.6 parrots/yr; Collazo et al. 2000, Collazo et al. 2010, White et al. 2014). Contributors to the sluggish population growth include raptor predation, ectoparasites, and environmental conditions associated with the rain forest (e.g., cavity microclimate; Snyder et al. 1987, FWS 1999).

By 2006, managers began reintroducing a second wild population within the Rio Abajo State Forest (hereafter Rio Abajo), from where the species had been extirpated since the 1930’s (Snyder, et al., 1987; USFWS, 2009; PRDNER, 2016). Reintroduced birds were originally sourced from the captive-breeding program. However, the population is currently comprised of both captive-bred released birds and their wild-born descendants (PRDNER, 2016). Currently, because of the impact of hurricanes, Irma and María only remain this wild population of approximately 80 individuals within the karst forest region of northcentral Puerto Rico, specifically the RAF (Snyder et al. 1987, Wunderle 1996, FWS 1999). The karst region of Puerto Rico was the last known occupied area outside CNF where parrots were abundant (Snyder et al. 1987). The karst region is drier than CNF and is currently the largest tract of contiguously forested habitat on the island (Chinea 1980, Snyder et al. 1987). Forests in this region contain many of the food plant species used by parrots (Chinea 1980, Cardona et al. 1986), and harbor a lower density of Red-tailed Hawks (0.00883 ± 0.0032/ha) than the Caribbean National Forest (0.01152 ± 0.0048/ha; Llerandi-Román 2006, Cardona et al. 1986, Frank Rivera-Milán unpubl. data, Collazo and Groom 2000) which had helped with the successful establishment of this wild population.

Despite the long history of the Puerto Rican Parrot Recovery Program (hereafter Recovery Program), only a handful of studies have been conducted on the species’ spatiotemporal dynamics. Furthermore, all the existing research on this topic has focused on the relic population within El Yunque (Snyder, et al., 1987; Lindsey, et al., 1991). Although early research included extensive habitat and diet characterization (Lugo & Frangi, 1993; Wunderle, 1999; Snyder, et al., 1987), El Yunque habitat differs greatly from existing and planned reintroduction sites on other parts of the island (Rivera & Aide, 1998; Ewel & Whitmore, 1973). Studies on other aspects of the species’ biology have been similarly restrictive in scope and location. The present study is the first to examine the spatiotemporal, social, and habitat use dynamics of the reintroduced population in Rio Abajo.

Recovery Program history

The Puerto Rican Parrot was once abundant throughout the entire island of Puerto Rico and found on at least three of the smaller neighboring islands (Snyder, et al., 1987). The decline of the species was precipitated by conversion of forest habitat to agricultural land during the 19th and 20th centuries, which reduced forest land to just 6% of total island cover (Birdsey & Weaver, 1987; Snyder, et al., 1987). By the 1940’s the species had been extirpated from most of its former range and the only extant population survived in the Luquillo mountain range, the island’s largest remaining area of forested land (Snyder, et al., 1987; Rodrigue-Vidal, 1959). A drastic population reduction to just 16 wild birds in the 1970’s prompted the creation of a captive breeding facility within El Yunque in 1973. A second captive breeding facility was established twenty years later within Rio Abajo to safeguard the species from stochastic events (USFWS, 2009). Reintroduction of wild birds in Rio Abajo commenced in 2006 and captive-born birds have been released into the wild almost every year since the program’s inception. Reintroduced birds have bred successfully in the wild since 2008. The Population increased by more than 20% per year through a combination of releases and breeding in the wild and reached its highest historical numbers in 2017 with more than 130 parrots (PRDNER, unpublished data).

Psittacine diet characteristics

Like many parrot species, Puerto Rican Parrots are mostly frugivorous but also eat other plant structures including seeds, flowers, leaves, twigs, bark, and leaf buds (Snyder, et al., 1987). Consumption of multiple plant parts is well documented among neotropical parrot species, although most species are primarily frugivorous or granivorous (Benavidez, et al., 2018). Larger parrot species generally consume a greater proportion of seeds in their diet, while smaller parrot species tend to consume more fruit (Benavidez, et al., 2018; Ragusa-Netto & Fecchio, 2006).

Seasonality in parrot diets is common since many tropical forests exhibit seasonal fruiting patterns (Wunderle, 1999; Lugo, 1993; Ragusa-Netto & Fecchio, 2006; Renton, et al., 2015). The highly variable temporal and spatial nature of fruit and seed abundance is believed to have resulted in plasticity in parrot foraging strategies (Renton, et al., 2015). One of the strategies employed by parrots is diet switching. Parrots respond to fluctuating food resources by switching between plant species or part consumed (Wermundsen, 1997; Ragusa-Netto & Fecchio, 2006; Renton, 2001; Renton, et al., 2015; Boyes & and Perrin, 2009). This can result high diversity in parrot diets. A review of 98 neotropical parrot species found that they consume 1293 plant species (Benavidez, et al., 2018). Puerto Rican Parrots in El Yunque have been documented feeding on more than 60 species of plants, many of which display seasonal or annual variation in fruiting (Snyder, et al., 1987; Rodrigue-Vidal, 1959; Wunderle, 1999).

Spatiotemporal dynamics

Parrot food resources also exhibit variation in spatial distribution and parrots may respond by shifting habitat use to exploit resources as they become available (Renton, et al., 2015). Puerto Rican parrots in El Yunque once exhibited seasonal movement across neighboring forest valleys (Rodrigue-Vidal, 1959). Anecdotal evidence also suggests that Puerto Rican parrots may have migrated from the Puerto Rico mainland to the neighboring island of Vieques during the rainy season (Wetmore, 1916). However, these movement patterns were lost as the species numbers continued to dwindle and the parrots’ range became more restricted (Snyder, et al., 1987).

Congeneric Lilac-crowned Parrots (Amazona finschi) shift foraging habitats between deciduous and semi-deciduous forests depending on available food resources (Renton, 2001). Parrots may also undertake short or long-distance migrations as a result of fluctuating food resources. Orange-bellied Parrots (Neophema chrysogaster) and Swift Parrots (Lathamus discolor) migrate between wintering and breeding grounds on Tasmania and mainland Australia (Drecsler, 1998; Mac Nally & Horrocks, 2000). Some Mealy Parrots (Amazona farinosa) engage in seasonal migrations spanning 200 km in what may be a response to seasonal abundance of fruit resources (Bjork, 2004). Direct evidence of habitat shifts in response to resource abundance exists for only a handful of parrot species. However, seasonal changes in abundance of many parrot species may be correlated with fluctuations in food resources and could indicate seasonal movements and resource tracking (Karubian, et al., 2005; Manning, et al., 2007; McFarland, 1991; Ragusa-Netto & Fecchio, 2006).

Home range

Historical home range estimates for Puerto Rican Parrots in El Yunque reflect the population decline throughout the years. The parrots occupied an estimated 3000 ha area within El Yunque during the mid-1950’s and occasionally made foraging flights outside the forest’s perimeter (Rodrigue-Vidal, 1959). Occupied habitat had dwindled to 2,200 ha by the 1970’s, although long-distance foraging flights of up to 7 km were still observed (Snyder, et al., 1987). Estimates of juvenile parrot home range in the 1990’s was only 1,372 ha with 5.5 km foraging flight distance (Lindsey, et al., 1991). However, the range contraction could be a result of differences between adult and juvenile spatial dynamics.

Home range sizes in other parrot species are heavily influenced by resource distributions. Congeneric Yellow-naped Amazons (Amazona auropilliata) that live in regions with dispersed vegetation range across an area ten times greater than individuals that live in regions with concentrated vegetation (Salinas-Melgoza, et al., 2013). As with habitat shifts, seasonal variations in home range size are believed to reflect seasonal changes in foraging resources. Kakapo parrots (Strigops habroptilus) have smaller winter home ranges, which may reflect the exploitation of locally abundant food resources (Walsh, et al., 2006).

Differences between juvenile and adult home range size have been documented for both Kakapo and Ground parrots (Pezoporus wallicus) (McFarland, 1991). Juveniles of both species have larger home ranges than adults and exhibit an exploratory shift during which they move away from one foraging area and return several days later (McFarland, 1991).

Behavior

Like other parrot species, Puerto Rican parrots exhibit bimodal activity patterns, with most of the physical activity taking place during the early morning and early evening hours. A typical parrot day begins with a period of intense vocalization which occurs close to sunrise, followed by foraging flights and then feeding within foraging areas (Snyder, et al., 1987). A small peak in feeding activity may be observed during midafternoon but most of the afternoon appears to be spent in a resting phase during which preening, stretching, and sleeping behaviors are commonly observed. Feeding and flight activity is again observed in the early evening hours followed by the return to roosting sites (Snyder, et al., 1987).

The bimodal activity pattern is common among neotropical parrot species and has also been observed in African, and Australasian parrot species (Boyes & Perrin, 2010; Legault, et al., 2012; Gilardi & Munn, 1998). Meyer’s parrots (Poicephalus meyeri) forage and fly in a bimodal daily pattern correlated with mean ambient temperature, an indication that physical activities are temporally distributed to avoid heat stress (Boyes & Perrin, 2010).

Sociality

Parrots are highly gregarious animals with strong social ties. Like other parrot species, the functional unit of Puerto Rican Parrot social systems is usually the mated-pair (Snyder, et al., 1987). Individuals of this species are monogamous and form stable pair bonds for the duration of their lives. Observations of flying parrots in El Yunque usually consisted of two or three birds. Larger groupings were rarely observed in flight together, although most foraging groups were comprised of more than three birds (Snyder, et al., 1987).

Outside of the mated-pair unit, parrots aggregate into larger foraging and roosting groups through fission-fusion dynamics (Hobson, et al., 2014; Buhrman-Deever, 2008). Group size and membership can vary over time. These characteristics are hypothesized to be related to the exchange of foraging information and the spatial distribution of resources (Chapman, et al., 1989). Foraging Brown-throated Conures (Aratinga pertinax) preferentially solicit overflying flocks by calling out to desired group members (Buhrman-Deever, 2008).

Legal Base

A legal framework needs to be considered when managing the critically endangered species. First, the constitution of the Commonwealth of Puerto Rico, Art. VI, sec. 19 establishes the following: “It will be the Commonwealth public policy the most efficient conservation of their natural resources, as so the best development and use of those resources for the benefit of the general citizens.

Under the second Article of the Law number 416 of 2004, Environmental Public Policy Law, is established that:

1. A public policy that promotes a desirable and convenient harmony between man and his environment;

2. Promote efforts that prohibit or eliminate damages to the environment and to the biosphere and promote the health and the wealth being of man;

3. Enrich the knowledge of the ecological systems and natural sources that are important to Puerto Rico.

In its third article, section (c), this law specifies that the government of the Commonwealth of Puerto Rico will follow a sustainable development by:

1. The most effective environmental and natural resources protection.

2. The most prudent and efficient use of the natural recourses for the benefit of its citizens.

The organic Law number 23 of 20 June 1972 created the Department of Natural Resources of Puerto Rico to become the executor of the natural resources of Puerto Rico. Law number 241 of 15 August 1999, also known as the New Wildlife Law of Puerto Rico, seeks to protect, conserve and promote the native and migratory wildlife species. Proclaim as property of Puerto Rico all the wildlife species in its jurisdiction and also define the faculties, powers and duties of the Secretary of the PRDNER., including the regulation of hunting activities, hunting weapons, and the inscriptions of those weapons. The Secretary also have the faculty of issue, renew and revoke hunting licenses and hunting permits for management and/or collection with scientific, educational, recuperation or population control purposes. Law number 241 establishes the regulations for the protection of the endangered of extinction species to Puerto Rico as well as the penalties for the violations of the dispositions of this law.

Regulation 6765 of the PRDNER define wildlife as any resident species whose natural propagation or survival does not depend of the watch, care or cultivation of man, it occurs naturally or has been adapted to the wild; any migratory species that visit Puerto Rico during any season of the year, and all exotic species established in the wild. The term wildlife includes birds, mammals, amphibians, invertebrates, and any part, product, nest, egg, hatch or dead body, and all threatened and endangered species. This Regulation tends to:

1. Promote the protection, conservation, and management of wildlife species.

2. Establish mechanisms to mitigate the natural habitat modifications.

3. Regulate rigorously the issue of hunting licenses, the hunting weapons inscriptions, and the revocation and suspension of the licenses because of infractions to the law and this regulation.

4. Regulate the introduction of exotic species to the territory of Puerto Rico.

5. Regulate all the activities related to the wildlife resources.

Regulation 6766 of the PRDNER tends to:

1. To identify, conserve and preserve the species listed as threatened and endangered of extinction;

2. Promote the population growth and survival of the threatened and endangered species;

3. To identify and promote the conservation of the critical natural habitats and essential critical natural habitats;

4. Regulate the import and export of the threatened and endangered species;

5. Adopt designation criteria used by the international scientific community to species with populations trends that can become endangered or even extinguish in a short period of time.

Finally, the Río Abajo State Forest is part of the Natural Patrimony Program under Law number 150 of 4 August 1988 as it is considered an area with extremely high natural and ecological values.

Study site

Puerto Rico is the smallest and easternmost island of the Greater Antilles (18°35’ N 65°37’W). It consists of the main island of Puerto Rico, the satellite islands of Culebra, Vieques and Mona, and is surrounded by numerous small islands and cays. The main island measures 178 km long and 58 km wide, with an area of 8,900 km2. Puerto Rico includes 6 ecological life zones (Ewel and Whitmore 1973); the three major life zones have been classified as subtropical dry, subtropical moist and subtropical wet, comprising approximately 98% of the total land area (Rivera-Milán 1992). Forest covers approximately 3400 km2 of Puerto Rico and is dominated by mature secondary forests (Ewel and Whitmore 1973). About 41.6% of the island is classified as closed canopy forest (Helmer et al. 2002). The three principal geographic regions of Puerto Rico are the montane-volcanic region of the Cordillera Central, a discontinuous fringe of mostly flat coastal plain and the rugged limestone or karst region occurring across the northern part of the island (Monroe 1976).

The limestone formations of Puerto Rico range in age from the early Cretaceous to the Quaternary, spanning some 146 million years. This region covers about 27.5% of the island’s surface and is subdivided into the northern, southern, and dispersed limestone areas (Lugo et al. 2001). It extends 140 km in an east-west direction along the north coast of Puerto Rico (Monroe 1976), is characterized by subtropical wet and moist life zones (Holdridge 1967), and approximately 88% of the region is located in the moist forest life zone (Lugo et al. 2001). The northern limestone region is dominated by highly eroded karst formations where elevations reach 530 m above sea level. This region contains Puerto Rico’s most extensive freshwater aquifer, largest continuous expanse of mature forest, largest coastal wetland estuary, and underground cave systems (Lugo et al. 2001).

The karst belt is characterized by closed depressions, subterranean drainages and caves, haystack hills (locally known as “mogotes”) and deep sink holes. The region constitutes about 142,544 ha (Monroe 1976). It is an extremely diverse region with unique flora and fauna, and many rare and endangered species. The flora of this region represents a transition between the wet forests of the Cordillera Central and dry forests of the coastal limestone (Lugo et al. 2001). This area is currently the largest contiguously forested region in Puerto Rico (Chinea 1980, Snyder et al.1987, Rivera and Aide 1998). Furthermore, karst forests contain the largest tree species richness in Puerto Rico (Lugo et al. 2001). Figueroa-Colón (1995) estimated the wet karst belt contains 23%, and the moist karst belt contains 16% of the endemic tree species of the island.

The Río Abajo State Forest, located within the moist limestone forest region of north-central Puerto Rico (18°20’N, 66°42’W), is the largest biological reserve found on limestone substrate in Puerto Rico (DNER 1976). This reserve is managed by the Department of Natural and Environmental Resources (DNER) Forestry Division. Río Abajo Forest is the largest reserve of the moist karst region and comprises an area of 2,300 ha with elevations ranging from 200-420 meters. Approximately 75% of the forest lies within the subtropical wet forest and the remaining 25% is within the moist life zone (DNER 1976). The area is characterized by seasonal rainfall patterns with the period of greatest precipitation between August and November. Annual precipitation averages 203.9 cm and annual temperature 24.8 °C (National Oceanic & Atmospheric Administration, NOAA). The closest NOAA weather centers to this area are the Dos Bocas and Arecibo Observatory stations. These recorded an average annual precipitation of 211.6 cm (2003) and 217.3 cm (2004), and a mean annual temperature of 25.3 °C during both years.

Due to agriculture and extractive practices, Río Abajo area was intensively deforested during the first three decades of the last century (Acevedo-Rodriguez and Axelrod 1999). The flora of Río Abajo is highly diverse and consists primarily of secondary forests in varying stages of succession. Alvarez-Ruiz et al. (1997) reported 242 tree species representing 51 families from moist and wet climates present in the forest. Forest plant communities vary from evergreen moist, mixed forests at the base of hills, to scrublands on dryer hill tops (Acevedo-Rodriguez and Axelrod 1999).

The fauna of this region is very important and highly diverse. A high degree of endemism exists, and includes 54% of all species known from Puerto Rico. Most of the native freshwater macrofauna, 73% of the island herpetofauna and 17 of the 18 endemic avian species of Puerto Rico occur in the karst belt region. Several species of migratory birds use this area as wintering habitat. The most common birds in this area are native or endemic species (Lugo et al. 2001). These include the Puerto Rican Ground Dove, Zenaida Dove, Ruddy Quail-Dove (Geotrygon montana), Bridled Quail-Dove (Geotrygon mystacea), Puerto Rican Tody (Todus mexicanus), Grey Kingbird (Tyrannus dominicensis), Pearly-eyed Thrasher (Margarops fuscatus), Puerto Rican Vireo (Vireo latimeri), Bananaquit, Black-faced Grassquit (Tiaris bicolor), Greater Antillean Grackle (Quiscalus niger), Puerto Rican Bullfinch (Loxigilla portoricensis), and the Puerto Rican Woodpecker (Melanerpes portoricensis).

Lugo et al. (2001) reported that frugivores are the most diverse and abundant guild in this region including pigeons and doves (Columbiformes), parrots (Psittaciformes), and a large diversity of songbirds (Passeriformes). Songbirds include the endemic Puerto Rican Bullfinch (Loxigilla portoricensis), and the Puerto Rican Stripe-headed Tanager (Spindalis portoricensis). These songbirds feed consistently on fruits and seeds of species commonly found in Río Abajo reserve (Acevedo-Rodriguez and Axelrod 1999, Aukema et al. 2005) such as palma real (Roystonea borinquena), ceboruquillo (Thouinia striata), moral (Cordia sulcata), yagrumo macho (Shefflera morototoni), yagrumo hembra (Cecropia schreberiana), cupey (Clusia rosea), and guaraguao (Guarea guidonia). These plant species represent more than 80% of frugivorous diet in this region (Collazo and Groom 2000). Some frugivorous birds are highly specialized in their diet. Saracco (2001) reported fourteen percent of foraging observations on G. guidonia were by the Puerto Rican Bullfinch in a study conducted in northcentral Puerto Rico. Although this species generally did not disperse seeds of G. guidonia, all trees visited by L. portoricensis were also visited by more effective dispersers. Moreover, the Pearly-eyed Thrasher and the Scaly-naped Pigeon (Columba squamosa) were detected more often in karst forests than in those of volcanic rock base (Rivera Milán 1993). Carlo Joglar (1999) found significant diet preferences among nine common frugivores in northcentral Puerto Rico. All bird species showed local preferences for a fruiting plant. Karst forests had lower fruit densities than shaded coffee plantations or moist forests outside of the karst belt.

More than 30 endangered or threatened species have been identified in the karst region. Nine endangered bird species have been reported, including two endangered raptor species, the Broad-winged Hawk (Buteo platypterus brunnescens) and the Sharp-shinned Hawk (Accipiter striatus vennator; Delannoy 1992, 1997). This area is also visited annually by thousands of Neotropical migrant birds. Forest passerines are represented by over 40 species, and 45 shorebird and seabird species (Raffaele 1992). The karst belt is spatially heterogeneous in terms of its landscape and climatic conditions. The subterranean and subaerial landforms promote a highly diverse ecological system. The diversity and abundance of wildlife in the karst belt is a result of its diversity of habitats and the karst topography, with its valleys, canyons, hills, sinkholes, and caves (Lugo et al. 2001). Furthermore, the locally extinct Puerto Rican Parrot (Amazona vittata) had been reported in this region.

Radio tagging

Captive parrots were selected for release based on genetic recommendations from the Recovery Program’s studbook, which contains pedigrees for the entire captive population (Earnhardt, et al., 2014). Parrots were kept in a T-shaped flight cage at the release site for a minimum of 6 months. The release cage was comprised of three adjoining sections that measured 6 m long X 6 m wide X 5.5 m tall each. Parrots could fly freely between adjoining cage sections. Parrots were fed a commercial pelleted diet. We subjected parrots to 6 months of pre-release conditioning routines, including flight conditioning and exposure to native foods (Appendix 1) (Collazo, et al., 2003). We also fitted parrots with dummy collars weighing 15 g (3-7 g heavier than actual transmitters) to acclimate them to transmitters.

We released 48 parrots over a period of three consecutive years (2015-2017). We released 15 parrots on January 21, 2015, 16 parrots on January 21, 2016, and 17 parrots on January 12, 2017. Released parrots were between 9 months to 4 years old. Most of the released parrots were bred at the Rio Abajo facility. However, six parrots in the 2015 group and three parrots in the 2016 group were bred at the El Yunque facility. Parrots were fitted with radio collars with unique frequencies between 164-165 MHz prior to release. We used Holohil (Carp, Ontario, Canada) SI-2C (12 g) transmitters for parrots in the 2015 release group (Meyers, 1996). For logistical reasons, we switched to ATS M1550 (6 g) (Isanti, Minnesota, USA) transmitters for parrots in the 2016 and 2017 release groups. Parrots were released at dawn to coincide with the most active period of other wild parrots. Supplemental food (a mix of commercial parrot pellets and bird seed) was provided throughout the year at 3-4 feeding stations located near the release cage.

Wild adult parrots

Wild birds were captured with the use of a baited trap-cage. The cage measured 0.9 m long X 1.2 m wide X 1.2 m tall and was suspended approximately 4.5 m in the canopy. The cage was placed close to existing supplemental feeding stations to encourage parrots to approach it. Modified feeding stations were placed inside the cage and filled with a combination of commercial parrot pellets and bird seed mix. The trap-cage was installed two to three months prior to the date of recapture to allow parrots to acclimate to it. Capture dates were decided once wild parrots showed signs of acclimation to the trap-cage (e.g. no hesitance at entry and 20 parrots entering at a time).

We captured 91 parrots and fitted 60 of them with radio collars during three nonconsecutive years (2013, 2015, and 2017). Transmitters were assigned to captured birds based on management priorities with emphasis being given to birds on which we had previously gathered little data. Between December 26-27, 2013, we captured a group of 36 parrots and fitted 21 with radio transmitters. On December 30, 2015, we captured 29 parrots and fitted 19 with radio transmitters. On December 27, 2017, we recaptured 26 parrots and fitted 20 with radio transmitters. We used Holohil SI-2C (12 g) transmitters for the 2013 capture group and ATS M1550 (6 g) transmitters for parrots in the 2015 and 2017 groups. Parrots were released back into the wild immediately after being fitted with a transmitter.

Juvenile wild parrots

Transmitters were attached to nestling parrots approximately 50 days after hatching. Nestlings were lowered to the ground for a veterinary exam prior to attaching the transmitters. We aimed to attach transmitters to as many wild nestlings as possible. If the number of nestlings exceeded the number of transmitters available, transmitters were randomly assigned to at least one nestling per nest. During the 2014 breeding season, we attached transmitters to 14 of 16 fledglings. During the 2015 breeding season, we attached transmitters to 20 of 25 fledglings. During the 2016 breeding season, we attached transmitters to 11 of 20 fledglings. During the 2017 breeding season, we attached transmitters to 11 of 31 fledglings. During the 2018 breeding season, we attached transmitters to 11 of 19 fledglings. We used Holohil SB-2C (6 g) transmitters for nestlings in 2014 and ATS M1550 (6 g) transmitters for nestlings between 2015-2018.

Radio tracking and home range

The home range of an animal is the space it uses in its regular activities (e.g., mating, searching for food, taking care of young, etc. (Burt, 1942). We used radio telemetry tracking techniques during this research to determine home range of on three radio-tagged focal parrot groups (i.e. newly released, wild adults, and wild juveniles). These techniques were necessary to obtain the most reliable information on parrots’s spatial ecology in the rough terrain of karst region. Terrain conditions in the karst area makes very difficult to assess population density, habitat use and spatial distribution by using other traditional methods. To date, telemetry is the most commonly used technique, not only because it estimates the size and shape of an animal’s home range (Petersen, 1979; Watson, 1986; Whitman, Ballard, & Gardner,1986), also because it enables the estimation of reproductive success, survival, movements, and interactions of individuals (Burger, Ryan, Jones & Wywialowski, 1991).

We tracked radio-marked parrots four to eight times monthly from January 2015 to December 2018 using R-1000 portable programmable receivers and flexible H-antenna RA-14K model. We located the animals by homing (i.e., animal terrestrially tracked) to get visual contact without altering its movement pattern. Before the tracking activities, we randomly determined the time of day (morning or afternoon) and the order in which individuals were daily tracked (Llerandi-Román, 2006). We tested the location error by taking compass bearings to transmitters placed on different forested area types and sinkholes and then comparing these bearings to known transmitter bearings (Llerandi-Román, 2006). After this test, we estimated overall observed errors of ±2º for these areas. For each visual location, we recorded weather condition, time, parrot’s activity (i.e., walking, perching, resting and foraging), habitat type and number of parrots in the area. Also, to estimate the position of radio-marked parrots we performed simultaneous triangulation. For triangulation, we determined parrots’ locations taking two simultaneous bearings from different positions. Triangulation allows location of poor-quality signals and reduces the average size of the confidence ellipse (Llerandi-Román, 2006). After locating a parrots, we did not relocate it during the next 12 hours to minimize serial correlation between successive locations (Llerandi-Román, 2006). We will use Locate II software (Nams, 1990) to generate UTM coordinates of parrots’ locations collected by triangulation. Using this program, we will determine the 95% error ellipses around locations generated by the triangulation method. To minimize the location error, we will not use the locations with error ellipses greater than the mean value of the core areas of the calculated home range for all marked individuals (Llerandi-Román 2006).

We monitored the animals equipped with VHF radio transmitters to determine their home range and to evaluate fluctuations in home range size with respect to habitats types associations and between breeding and non-breeding season. We will use the saturation curve method to determine the location sample necessary to calculate a home range size with the minimal bias. We will plot each VHF location recorded from each radio tagged parrot to generate fixed kernel home range (95%) and core areas (50%), using the least squares cross validation smoothing factor (Seaman & Powel, l996). We will estimate temporal (i.e., breeding vs non-breeding) and annual home range size of each parrot using Animal Movement Extension (Hooge & Eichenlaub, 1997) in Arc Map 10.6, We will perform tests of normality on each data set. When data did not violate the normality assumptions we will use sample Student t-tests (when compare two groups) and Analysis of Variance (ANOVA; when compare more than two groups) to evaluate if home range and core area size of radio tagged parrot differed due to gender and between focal parrot groups (i.e. newly released, wild adults, and wild juveniles). If data violated the normality assumption a Mann-Whitney Rank Sum test (instead of the Student T-test) will be performed. The same set of statistical analyses will be use to evaluate if the radio tagged parrots home range and core area size differed between seasons (i.e. breeding and non-breeding). In all analyses statistical significance will be considered when P ? 0.05.

Habitat use

For habitat type analysis, we will use the land cover map of the Río Abajo Stae Forest and adjacent areas generated by Martinuzzi et al. (2008). Using Analysis Tools Extension in Arc Map 10.6 , we will clip the habitat types represented within the home range of each VHF radio-tagged parrot. Then, we will calculate the proportion of each habitat type within every parrot’s home range. We will use euclidean distances (Conner & Plowman 2001, Conner et al. 2003) within Johnson’s (1980) hierarchical resource selection framework to test if parrots use habitats in a random (nonpreference) or non-random (preference) way. This procedure considers each individual as the sampling unit, does not require assigning each radio marked individual to a certain habitat type, and is robust to location error (Conner & Plowman 2001, Llerandi –Román, 2006). Because of small relocation sample size, we will use parrots’ annual home range (95% fixed kernel) as a framework to conduct a hierarchical resource selection of third order analysis (i.e. determine the use of the different habitat types within the parrots’ home range). Using the Animal Movement extension (Hooge & Eichenlaub, 1999) in Arc Map 10.6, we will generate the same quantity of random points (expected) as animal locations collected (observed) from a uniform distribution within the annual home range of each collared parrot. From these random points, we will calculate a vector of mean distances (ri) to the nearest habitat type using Analysis Tools Extension in Arc Map 10.6, We will repeat this procedure using the locations of each parrot to produce a vector of mean distances to the nearest habitat type (ui). We will create a vector of ratios (di) for each animal by dividing each element (value for habitat type) in the animal mean distance vector (ui) by the corresponding element in the random mean distance vector (ri). If the habitat is used randomly, the distance of animal (ith) locations to each habitat (ai) should equal the distance between random points and each habitat (ri); therefore the ai:ri distance ratio should equal 1.0. If use was nonrandom, the distance ratio ai:ri should be less or greater than 1.0. Element values (?) of the ratio vector < 1.0 indicate locations closer than expected to a habitat (selection), whereas element values > 1.0 indicate the habitat was used less than expected (avoidance) (Llerandi-Román, 2006). We will use a Multivariate Analysis of Variance (MANOVA, Sokal & Rohlf, 1995) to evaluate non-random use of habitat types. If a non-random use was detected a t-test will be conducted to determine which habitats were used unequally (Conner & Plowman, 2001).

Survival

To estimate the monthly survival of the VHF radio tagged parrots from January 2015 to December 2018, we will use the Known-fate model in Program MARK 3.2. The Known-fate model assumes that all marked animals are alive at the beginning of the interval, each animal has the same survival probability, animal’s fate is known with certainty through the interval, and censoring of marked individuals is independent of their fate (White & Burnham, 1999).

Since radio marked parrots were sampled at different time periods during the study, we will use a staggered entry design (Pollock et al., 1989) to model parrot’s monthly survival (Llerandi-Román, 2006). This approach assumes that animals are marked, radio-collars do not affect survival probability, and newly marked individuals are assumed to have the same survival function as previously marked ones (Llerandi-Román, 2006). To test if the covariates time specific variation, gender, and focal parrot groups (i.e. newly released, wild adults, and wild juveniles) affect the monthly survival we will use the program MARK to develop the following six analysis models: {S(.)}, {S (Focal Group)}, {S(Sex)}, {S (Sex+Focal Group)}, {S(Sex+T)}, {S(t)}. We will use the Akaike’s Information Criterion (AIC) to determine the model or models that best fit theradio tagged parrot’s monthly survival data (Akaike 1973; Burnham & Anderson, 1998; Dinsmore et al., 2002; Llerandi-Román, 2006).

Diet characteristics

The diet of Puerto Rican Parrots throughout the year will be determined by observations of feeding activity. Data will be collected during each tracking activity. Whenever we track an individual or a group of Puerto Rican Parrots and found them feeding, we will record the name of the plant species, the part of the plant the parakeets were consuming, the time, date and place where parrots were observed and, when possible, the group size. An observation of a group of parrots on a plant species will be recorded as a feeding bout (Altmann 1974, Cannon 1981, Galetti 1993, O’Donnell& Dilks 1993, Wermundsen 1997)).

The results will be based on the frequency of feeding bouts. Using this method, it is possible to determine the food sources consumed by the parrots. Likewise, the period when each feeding source is available will be determined. The diversities of the

Did you like this example?

592

31