Contributed Paper

Two Hundred Years of Local Avian Extinctions in Eastern Amazonia ´ RGILA G. MOURA,∗ ALEXANDER C. LEES,†‡‡ ALEXANDRE ALEIXO,† JOS BARLOW,†‡ NA ´ TIMA C. LIMA,† SIDNEI M. DANTAS,† JOICE FERREIRA,§ MARIA DE FA ∗∗ AND TOBY A. GARDNER¶ †† ∗

Curso de P´ os-Graduac¸˜ao de Zoologia, Universidade Federal do Par´a/Museu Paraense Em´ılio Goeldi, Caixa Postal 399, CEP 66040-170, Bel´em, Par´a, Brazil †Coordenac¸˜ao de Zoologia, Museu Paraense Em´ılio Goeldi, Caixa Postal 399, CEP 66040-170, Bel´em, Par´a, Brazil ‡Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ, United Kingdom §Embrapa Amazˆ onia Oriental, Trav. Dr. En´eas Pinheiro s/n, CP 48, CEP 66095–100 Bel´em, PA, Brazil ¶Department of Zoology, University of Cambridge, Cambridge, CB2 3EJ, United Kingdom ∗∗ International Institute for Sustainability, Rio de Janeiro, CEP 22460-320, Brazil ††Stockholm Environment Institute, 87D Linegatan, Stockholm, Sweden

Abstract: Local, regional, and global extinctions caused by habitat loss, degradation, and fragmentation have been widely reported for the tropics. The patterns and drivers of this loss of species are now increasingly well known in Amazonia, but there remains a significant gap in understanding of long-term trends in species persistence and extinction in anthropogenic landscapes. Such a historical perspective is critical for understanding the status and trends of extant biodiversity as well as for identifying priorities to halt further losses. Using extensive historical data sets of specimen records and results of contemporary surveys, we searched for evidence of local extinctions of a terra firma rainforest avifauna over 200 years in a 2500 km2 eastern Amazonian region around the Brazilian city of Bel´em. This region has the longest history of ornithological fieldwork in the entire Amazon basin and lies in the highly threatened Bel´em Centre of Endemism. We also compared our historically inferred extinction events with extensive data on species occurrences in a sample of catchments in a nearby municipality (Paragominas) that encompass a gradient of past forest loss. We found evidence for the possible extinction of 47 species (14% of the regional species pool) that were unreported from 1980 to 2013 (80% last recorded between 1900 and 1980). Seventeen species appear on the International Union for Conservation of Nature Red List, and many of these are large-bodied. The species lost from the region immediately around Bel´em are similar to those which are currently restricted to well-forested catchments in Paragominas. Although we anticipate the future rediscovery or recolonization of some species inferred to be extinct by our calculations, we also expect that there are likely to be additional local extinctions, not reported here, given the ongoing loss and degradation of remaining areas of native vegetation across eastern Amazonia.

Keywords: avifauna, colonization, forest dependency, IUCN Red List, sighting record, urban conservation Doscientos A˜ nos de Extinciones Locales de Aves en la Amazonia Oriental

Resumen: Las extinciones locales, regionales y globales causadas por la p´erdida, degradaci´on y fragmentaci´ on de h´ abitat se han reportado ampliamente para los tr´ opicos. Los patrones y conductores de esta p´erdida de especies son ahora muy conocidos en la Amazonia, pero a´ un persiste un vac´ıo significativo en el entendimiento de las tendencias a largo plazo en la persistencia de las especies y la extinci´ on en terrenos antropog´enicos. Dicha perspectiva hist´ orica es cr´ıtica para entender el estado y las tendencias de la biodiversidad existente as´ı como para identificar prioridades que detengan las p´erdidas futuras. Al usar

‡‡ Address

correspondence to A. C. Lees, email [email protected] Paper submitted October 30, 2013; revised manuscript accepted January 26, 2014.

1 Conservation Biology, Volume 00, No. 0, 1–11  C 2014 Society for Conservation Biology DOI: 10.1111/cobi.12300

2

Avian Extinctions in Amazonia

juegos de datos hist´ oricos extensivos de registros de espec´ımenes y resultados de sondeos contempor´ aneos, buscamos evidencias de extinciones locales de avifauna en una selva de tierra firme a lo largo de 200 a˜ nos na de Bel´em. Esta regi´ on en una regi´ on de 2, 500 km2 en la Amazonia oriental, cerca de la ciudad brasile˜ tiene la mayor historia de trabajo de campo ornitol´ ogico en toda la cuenca del Amazonas y yace en el altamente amenazado Centro de Endemismo de Bel´em. Tambi´en comparamos nuestros eventos de extinci´ on inferidos hist´ oricamente con los datos extensos de la ocurrencia de especies en una muestra de zonas de influencia en una municipalidad cercana (Paragominas) que engloba un gradiente de p´erdidas pasadas de bosque. Encontramos evidencias de la extinci´ on posible de 47 especies (14% del acervo regional de especies) que no se reportaron de 1980 a 2013 (80% fue reportado por u ´ ltima vez entre 1900 y 1980). Diecisiete especies aparecen en la Lista Roja de la UICN, y muchas de estas son de gran tama˜ no. Las especies perdidas en la regi´ on inmediata a Bel´em son similares a aquellas que actualmente est´ an restringidas a zonas de influencia con bosques bien desarrollados en Paragominas. Aunque anticipamos el futuro redescubrimiento o recolonizaci´ on de algunas especies que se infiere est´ an extintas de acuerdo a nuestros c´ alculos, tambi´en esperamos que probablemente existan extinciones locales adicionales, no reportadas aqu´ı, dada la p´erdida y degradaci´ on continua de las a on nativa a lo largo de la Amazonia oriental. ´ reas restantes de vegetaci´

Palabras Clave: avifauna, colonizaci´on, conservaci´on urbana, dependencia de bosque, Lista Roja de la UICN, registro de avistamientos

Introduction Determining if and when a species becomes locally, regionally, or globally extinct is never easy because the rarer a species gets, the more difficult it is to find and study. The ecological literature contains many cases whereby species once considered extinct have been rediscovered decades or even centuries later (e.g., Crowley 2011). Such cases indicate the necessity for thorough field surveys before declaring a species to be extinct (e.g., Butchart et al. 2006). These problems are most severe in the tropics, where biodiversity is richest, extinction (and rediscovery) rates are highest (Costello et al. 2013), and land-use change is currently most acute (e.g., Sodhi et al. 2004). Furthermore, a persistent lack of time-series data means that estimating extinction risk is often only possible through a space-for-time substitution. For example, one might compare current species persistence in heavily forested areas with patterns of persistence in adjacent recently deforested areas (Pickett 1989). However, it is possible that many short-term studies may not be able to capture long-term patterns of species survival because extinction debts may be paid over a very long period for species with long generation times and insulation from real-world threats (in experimental landscapes or over short periods), such as hunting and trapping, may lead to underestimates of extinction risk (e.g., Peres et al. 2010). Thus, understanding what constitutes the baseline of biodiversity prior to recent large-scale landscape change and resource extraction by humans is fraught with difficultly for many parts of the world. One of the most robust approaches to estimating past extinctions and current extinction risk is to combine data from current field surveys with historical data from museum specimens (e.g., Kattan et al. 1994; Patten et al. 2010). This powerful approach is only possible for regions of the world where extensive natural history field-

Conservation Biology Volume 00, No. 0, 2014

work has been conducted over long periods. Here, we present such a case study for the region around the city of Bel´em, the longest studied area of the entire Amazon basin. Amazonia is subject to the highest absolute rates of loss of tropical forest on the planet (Hansen et al. 2013). Despite substantial reductions in deforestation rates in the Brazilian Amazon during the last decade (INPE 2013), this loss and fragmentation of forests threaten many species with global extinction (e.g., Bird et al. 2011). Local extinctions caused by habitat loss, fragmentation, and disturbance from fire and logging are relatively well studied in Amazonia (Peres et al. 2010), and the longest running study, the Biological Dynamics of Forest Fragments Project, dates back only to 1979 (Laurance et al. 2011). Thus, a major knowledge gap relating to the conservation of Amazonian biota is an understanding of longterm trends in species persistence and extinction in the human-modified landscapes that increasingly characterize the region. Arguably the most threatened region of Amazonia is the 243,000 km2 Bel´em Center of Endemism (BCE) in northeastern Par´a and W Maranh˜ao states. This area has the longest history and highest proportion of forest loss of any Amazonian interfluve and retains 24% of its original primary forest cover (Almeida & Vieira 2010). Most of this remaining forest is heavily fragmented and degraded by recurrent selective-logging and fire events, as well as overextraction of game animals and other nontimber forest resources (Almeida & Vieira 2010; Amaral et al. 2012). Although 11 of the 160 threatened avian taxa on the current Brazilian Red List are strictly Amazonian, 10 of these are restricted to the BCE (Machado et al. 2008). A further 12 species occurring in but not necessarily restricted to the BCE are also considered globally threatened (IUCN 2013), of which 7 do not appear on the Brazilian Red List (Machado et al. 2008). Quantitative data on the responses

Moura et al.

3

70°0'W

60°0'W

50°0'W

40°0'W

0

250

500 Km

0°0' 0°0'

1. 2.

10°0'S 10°0'S

20°0'S 20°0'S

a)

30°0'S

b)

0

5

10 Km

c)

30°0'S 70°0'W

60°0'W

50°0'W

40°0'W

30°0'W

Figure 1. (a) The state of Par´ a in Brazil, (b) position of the study areas (1, metropolitan region of Bel´em (MRB); 2, Paragominas in Par´ a), and (c) the extent of remaining forest cover in the MRB (black, forest; dark gray, secondary forest; light gray, urban; white agricultural; hatched, water).

to land-use change and persistence in fragmented landscapes for many of these threatened taxa are generally lacking (Portes et al. 2011). The state capital Bel´em (Bel´em do Gr˜ao Par´a) was founded by the Portuguese in 1616, and by 1752 it was a hub of Amazonian biodiversity research (Teixeira et al. 2010). Many renowned naturalists (including Henry Walter Bates, Alfred Russel Wallace, and Johann Baptist von Spix among others) collected thousands of bird (among other) specimens that were deposited in museums across the world (Novaes & Lima 2009). This wealth of specimen data has, until now, been untapped for use in conservation research, yet offers invaluable insights into the history of local avifaunal extinctions in this unique corner of Amazonia (e.g., Burgman et al. 1995). In stark contrast to most of Amazonia, where substantial forest losses have only occurred in the last 50 years, the Metropolitan region of Bel´em (MRB) was already heavily deforested and defaunated more than a century ago (Vieira et al. 2007). Our goal was thus to investigate long-term trends in local extinction and persistence of Amazonian birds in a highly fragmented and degraded forest region (MRB) that characterizes the BCE. We searched for evidence of local extinctions from 1812 to 1980 in the MRB by examining museum records and estimating the persistence probabilities and likely extinction dates for species unrecorded since 1980. We compared these historical data with field data on local extinctions from the most extensively forested region left in the BCE—the municipality of Paragominas, which was subject to a recent exhaustive biodiversity inventory (Lees et al. 2012; Gardner et al. 2013; Moura et al. 2013). We then determined if current patterns of threat recognized at the regional, national, and international levels agreed with our findings of which species were most threatened with extinction in both our historical and contemporary data sets.

Methods Study Site The 2537 km2 MRB (approximately 1°S, 48°ʹW, Fig. 1), a conurbation of 2.2 million people, is in northeastern Brazil at the mouth of the Guam´a River in Par´a state and is divided into 6 municipalities: Bel´em, Ananindeua, Marituba, Benevides, Santa Isabel, and Santa B´arbara do Par´a. Local soils are typically deep oxisoils, well-drained with low natural fertility, and the natural climax vegetation is dense ombrophilous terra firma and v´ arzea (seasonally flooded) forests (Novaes & Lima 2009). Local climate details are in Supporting Information. Since its founding on 12 January 1616 the city and its environs have undergone over 4 centuries of forest loss and degradation, catalyzed by the construction of the Estrada Real from 1616 onward, a highway that connects Bel´em to S˜ao Lu´ıs in Maranh˜ao state, and later associated with the construction of the Braganc¸a railway, between 1883 and 1908 aimed at developing eastern Par´a (Vieira et al. 2007). Present day forest cover in the MRB is approximately 43.5% of the original pre-Columbian extent (INPE 2013). There are 6 protected areas (6399 km2 ), most of which are severely degraded due to ongoing disturbance events such as fire, selective-logging, and illegal hunting (Le˜ao et al. 2007).

Data Collection Novaes and Lima (2009) list 490 bird species for the MRB, of which 329 (67%) are present in unflooded terra firma forest. We compiled a baseline list of those species reliably recorded in the region and represented either by specimen records or archived documented observations (voucher images or sound recordings). The principle

Conservation Biology Volume 00, No. 0, 2014

Avian Extinctions in Amazonia

4

source of specimen data came from digitized records of study skins deposited at the Museu Paraense Em´ılio Goeldi (MPEG) and from 9 North American institutions with data from the MRB archived on the ORNIS database (www.ornisnet.org). The databases of many avian collections are undigitized, so we also conducted an exhaustive search of secondary sources to locate additional historical records (e.g., Sclater & Salvin 1867; Hellmayr 1905; Stone 1928) from which we extracted collection dates directly or used to direct targeted visits to other collections (e.g., the Natural History Museum at Tring). To ascertain the contemporary presence of species in the region, in addition to specimen data we also accepted digital voucher images and sound recordings (Supporting Information). Through the compilation of data from multiple sources, we compiled a sighting record for each species. A full compilation of our data sources is listed in Supporting Information.

Data Analyses We considered 1980, the cut-off year for analysis. Species unrecorded since 1980 were candidates for local extinction because the last really rigorous inventories were finished in the 1970s (e.g., Lovejoy 1971; Novaes 1973). To infer extinction dates, we used the formulas of Solow (1993, 2005). These formulas assume a uniform distribution of sightings (nonstationary Poisson, which also assumes the records are independent). We organized the records as t1 < t2 < . . . < tn , where n represents the number of times a species was sighted during the study period, ordered from the earliest to latest, starting with t1 = 0. To determine if a species was likely to be extinct, we used the formula p = (tn/T) n−1 , where T is the difference between the first sighting and the target year 2013 (which is the endpoint of the study period that corresponds to the present time). If p < 0.05, then the species was considered likely to be extinct (Solow 1993, 2005). We used the following equation to estimate extinction dates:TˆE = (n + 1/n)tn, where the expected year of extinction is TˆE and the first record is t1 (for species with at least 4 independent records). The confidence interval for TˆE was calculated as TE U = tn /α 1/n , where α = 0.05. Because the Solow equation is heavily dependent on the initial number of sightings (tn ), calculating the sighting rate (sensu McInerny et al. 2006) is also useful because it allows for greater comparability between taxa if there is disparity in the period of initial sighting or discovery. The sighting rate is calculated with p = (1 − (n/tn)(T −tn) ). We present both measures for greater confidence in our inferred candidates for extinction. To assess the generality of extinction patterns, we compared them with the results of a 2010–2011 survey investigating patterns of avian persistence in 18 catchments (of approximately 5000 ha) in the municipality of Paragomi-

Conservation Biology Volume 00, No. 0, 2014

nas (2°S, 47°W), which is 307 km south of the MRB and has a baseline avifauna that is nearly identical to that of the MRB (Lees et al. 2012; Gardner et al. 2013). The 18 catchments studied were delineated using a digital elevation model and SWAT (Soil and Water Assessment Tool) for ARCGIS 9.3 and represented a gradient of accumulated forest loss (based on classified 2010 Landsat images [Gardner et al. 2013]) from 94% (6% remaining primary forest cover) to 0% (100% forest cover). Between 7 and 12 (300 m) transects were stratified (forest, nonforest) across each catchment. Three point count stations were allocated to each transect (see Lees et al. [2012] and Gardner et al. [2013] for more details on experimental design and avian survey protocols). We compiled a list of avian taxa classified as threatened on the state (Par´a) (Aleixo 2006), national (Brazilian) (Machado et al. 2008), and international (IUCN) red lists (IUCN 2013) (evaluations at the state and national level also considered taxa below the species level, in some cases pending taxonomic upgrades) currently or historically occurring in the MRB. We then compared all species that we inferred to be extinct in the MRB and all species listed as threatened on state, national, and international lists with the total number of records, catchment occupancy, land-use breadth (number of land uses sampled from primary and secondary forest, pasture, silviculture, and mechanized agriculture), and the minimum percentage of forest cover within each occupied catchment in Paragominas. To assess potential traits of birds vulnerable to local extinction, we compared the threats compiled for each species and species’ life history attributes, such as mass and forest dependency, by Birdlife International (2013). Finally, we adapted the framework of Butchart et al. (2006) and used it to assess the conservation status of species either already listed or deemed to be regionally threatened based on our analyses of species persistence in the MRB. For species last recorded quite recently, there needs to be greater confidence that the last individual has died before the species can be considered extinct.

Results Possible Extinctions in the MRB We traced 10,147 specimens of 329 terra firma species from 8 collections from which we were able to construct a sighting record. These indicated that 47 terra firma species (14%) were unrecorded in the MRB over the last 33 years, a loss of 0.28 species/year. Probable extinctions were inferred to have occurred over the course of the entire period 1800–2000. However, more occurred in the 20th century; 80% of candidate extinctions were recorded between 1900 and 1980 (Table 1). We were able to estimate extinction probabilities for 26 (for which number of records was ࣙ4). Ten species were

Moura et al.

5

Table 1. Species unrecorded in the last 33 years in the metropolitan region of Bel´em. Scientific namea

English common name

Penelope pileata Aburria cujubi Pauxi tuberosa Odontophorus gujanensis Sarcoramphus papa Accipiter poliogaster Accipter superciliosus Harpia harpyja Spizaetus melanoleucus Spizaetus ornatus Daptrius ater Micrastur mirandollei Micrastur semitorquatus Psophia obscura Anodorhynchus hyacinthinus Ara macao Ara chloropterus Guarouba guarouba Pyrrhura lepida Touit huetii Deroptyus accipitrinus Neomorphus geoffroyi Lophornis gouldii Calliphlox amethystina Piculus chrysochloros Celeus torquatus Hylopezus macularius Sclerurus caudacutus Sclerurus rufigularis Dendrocincla merula Deconychura longicauda Dendrexetastes rufigula Berlespchia rikeri Philydor ruficaudatum Philydor erythrocercum Myiobius atricaudus Phoenicircus carnifex Gymnoderus foetidus Platyrinchus platyrhynchos Platyrinchus saturatus Piprites chloris Corythopis torquatus Hylophilus ochraceiceps Lamprospiza melanoleuca Dacnis lineata Periporphyrus erythromelas Euphonia minuta

White-crested Guan Red-throated Piping-Guan Razor-billed Curassow Marbled Wood-Quail King Vulture Gray-bellied Hawk Tiny Hawk Harpy Eagle Black-and-white Hawk-Eagle Ornate Hawk-Eagle Black Caracara Slaty-backed Forest-Falcon Collared Forest-Falcon Dark-winged Trumpeter Hyacinth Macaw Scarlet Macaw Red-and-green Macaw Golden Parakeet Pearly Parakeet Scarlet-shouldered Parrotlet Red-fan Parrot Rufous-vented Ground-Cuckoo Dot-eared Coquette Amethyst Woodstar Golden-green Woodpecker Ringed Woodpecker Spotted Antpitta Black-tailed Leaftosser Short-billed Leaftosser White-chinned Woodcreeper Long-tailed Woodcreeper Cinnamon-throated Woodcreeper Point-tailed Palmcreeper Rufous-tailed Foliage-gleaner Rufous-rumped Foliage-gleaner Black-tailed Flycatcher Guianan Red-Cotinga Bare-necked Fruitcrow White-crested Spadebill Cinnamon-crested Spadebill Wing-barred Piprites Ringed Antpipit Tawny-crowned Greenlet Red-billed Pied Tanager Black-faced Dacnis Red-and-Black Grosbeak White-vented Euphonia

a Our taxonomy follows the checklist of b Number of records. c The interval between the first and last d The time interval between the e The p values shown are based

Last record nb 1870 1835 1959 1906 1898 1915 1968 1894 1962 1812 1848 1835 1968 1922 1812 1900 1906 1848 1968 1969 1912 1912 1967 1926 1963 1968 1968 1967 1972 1968 1968 1936 1935 1968 1965 1968 1930 1912 1968 1972 1965 1972 1969 1965 1926 1905 1960

2 1 6 7 1 3 6 2 1 1 1 1 2 9 1 4 6 3 4 4 4 4 9 1 6 1 6 5 18 6 3 1 4 1 6 6 10 1 9 12 2 11 8 8 6 1 3

tn c

Td

Solowe

124 71 133 87 88 94 156 71 100 100 55 62 133 42 60 5 34 86 96 72 56 60 124 66 130 47 -

178 178 178 178 201 104 201 115 201 201 101 112 178 88 101 50 112 134 141 95 101 101 165 110 178 134 -

0.164 0.004 0.233 0.003 0.084 0.523 0.468 0.235 0.123 0.123 0.008 0.052 0.233 0.052 0.000 0.00 0.028 0.109 0.146 0.083 0.009 0.003 0.057 0.028 0.111 0.005 -

Estimated Upper Sighting Mean extinction year 95% bound rate weight (g) 1918 1931 1974 1976 1969 1946 1975 1977 1978 1935 -

1952 1956 1992 1984 1972 1993 1993 1981 2004 1965 -

0.108 0.991 0.178 0.000 0.020 0.491 0.417 0.150 0.046 0.046 0.001 0.015 0.178 0.010 0.000 0.00 0.001 0.000 0.090 0.046 0.001 0.000 0.032 0.007 0.070 0.000 -

1260 1407 2813 314 3337 420 104 6550 751 1215 352 517 631 1071 1565 1245 1479 260 75 60 255 363 2 2 66 124 40 38 22 53 26 70 35 27 25 10 84 283 13 13 19 17 15 35 11 48 9

Brazilian birds compiled by the Comitˆe Brasileiro de Registros Ornitol´ ogicos (CBRO 2011).

record. year of the first sighting and the target year (2013). on the Solow equation (Solow 1993, 2005).

considered likely to be extinct based on the Solow equation; extinction dates ranged from 1918 to 1978 (95% confidence intervals between 1952 and 2004, Fig. 2), and 16 species were presumed to be still extant. Of the 21 species for which it was not possible to calculate the extinction probability, 9 were unrecorded after 1900. We were reasonably confident that these species have become locally extinct (Fig. 2) because all are susceptible to local extinction following habitat loss and hunting (e.g., Peres 2000) and all are large-bodied (> 350 g) easy to survey species which are very unlikely to have remained undetected for over a century. The sighting rate calculations for 26 species suggested that 17 species were likely to be extinct in the MRB, including 8 species that were

highly likely to be extant based on the Solow equation (Table 1).

Forest Dependency of Threatened Species In comparing our historical analysis with contemporary avian surveys in the municipality of Paragominas, we found that 8 species recorded as extinct (since 1980) in the MRB were also unrecorded in Paragominas. Thirtynine species considered extinct in the MRB persisted in Paragominas, although 5 of these are likely to be rare given that they were not recorded during our comprehensive 2010 survey and reported only by Portes et al.

Conservation Biology Volume 00, No. 0, 2014

6

Avian Extinctions in Amazonia

Figure 2. Last record, extinction date (year), and confidence intervals for 47 species unrecorded in the Metropolitan region of Bel´em, Brazil, after 1980 (gray squares, last record of species presumed extinct but for which there are insufficient records to use the Solow equation (n < 4); white triangles, last record for species considered extant based on Solow equation; black circles, date of last record for species considered extinct based on Solow equation; white circles, extinction date inferred based on Solow equation; half-filled diamond, species extant based on Solow equation and extinct based on sighting rate; half-filled circle, species extinct based on Solow equation and extant based on sighting rate; crosses, 95% confidence interval for extinct species based on Solow equation).

(2011). Eight of 11 threatened (from any list) species that were still extant in the MRB were also recorded in Paragominas (Supporting Information). Forest dependency of individual threatened or extinct species in Paragominas (measured as the minimum percentage forest cover at the catchment scale from which a species was recorded) ranged from 34% to 100%; the

Conservation Biology Volume 00, No. 0, 2014

number of catchments occupied ranged from 1 to 15 (of 18 catchments surveyed). The number of catchments occupied in Paragominas and the date of the last record in the MRB were weakly and positively related (Fig. 3, nonlinear exponential model r2 = 0.16; p < 0.05). Species recorded in fewer catchments were more susceptible to local extinction (last recorded longer ago). Among

Moura et al.

7

in old secondary forests [>50 years] elsewhere in the MRB). Nine other species were also recorded in secondary forests, and 4 species considered likely to be extinct in the MRB and listed as endangered on the IUCN list were also threatened by hunting and the wild bird trade. Thirteen of the 15 species last recorded before 1900 were either large-bodied (>1 kg) game birds, raptors, or psittacids of high commercial value (Supporting Information). A combined analysis of sighting records, threat status, and estimates of the ease of detection revealed 15 extinct species for which we are very confident of local extinction, 19 probably extinct species, and 13 possibly extinct species for which we have lower confidence in their extinction (Fig. 4). Figure 3. Relationship between inferred extinction date (year) in the Metropolitan region of Bel´em and number of catchments occupied in Paragominas, Brazil (r2 = 0.16, p < 0.05).

the 47 species with a high extinction probability in the MRB, 9 were recorded once in Paragominas, indicating parallel patterns of rarity. However, this was not the case for all species. For example, the Red-billed Pied Tanager (Lamprospiza melanoleuca) was recorded on 37 occasions from 12 catchments in Paragominas (Supporting Information). Reassessment of Species Threat Status Fifteen (31%) of the 47 species unrecorded since 1980 are categorized as threatened; 10 of these species are on the Par´a list, 6 are on the Brazilian list, and 10 are on the IUCN list (Supporting Information). Ten species listed as regionally threatened are still extant in the MRB (Supporting Information). All 47 species we considered to have a high probability of extinction in the MRB, together with the species currently considered to be endangered in Par´a (12 species), are threatened by habitat loss, degradation, and fragmentation, and 19 of these species were only recorded from primary forests in Paragominas, of which 3 species were found exclusively in remnant fragments of undisturbed forests (Dark-winged Trumpeter [Psophia obscura], Guianan Red-Cotinga [Phoenicircus carnifex], and Tawny-crowned Greenlet [Hylophilus ochraceiceps rubrifrons]). Fifteen species occupied both primary and secondary forest and 3—Scarlet Macaw (Ara macao), Red-and-green Macaw (Ara chloropterus), and Golden Parakeet (Guarouba guarouba)—were also recorded in nonforest (agricultural) areas (Fig. 3). Of the species considered threatened in Par´a, only Black-spotted Bare-eye (Phlegopsis nigromaculata paraensis) was restricted to primary forest around Paragominas (although it occurs

Discussion We found evidence for the possible local extinction of 47 terra firma species in the MRB since 1812, all of which remained undetected from 1980 to 2013. Gradual local and regional extinctions have been reported throughout the Neotropics (Table 2), but we conducted the first study illustrating the long-term (>160 years) erosion of an Amazonian bird community. Rate of loss was estimated at 0.28 species/year in the MRB. By comparison, Robinson (2001) found that Barro Colorado Island, Panama, lost 13.5% of its avifauna in 25 years (1.1 species extinction/year) and Patten et al. (2010) found that Palenque, Mexico, lost 9.5% of its avifauna in 109 years (0.21 extinctions/year). These results suggest that short-term studies of avian extinctions from Amazonian forest landscapes may yield very conservative results because extinction debts may be paid over a long period and species present in the current landscape may not be part of viable populations (e.g., Brooks et al. 1999; Metzger et al. 2009). Our results therefore reinforce the critical importance of establishing (when historical records are available) an accurate local baseline for a given biota to avoid underestimating levels of species losses associated with cumulative landuse change and synergistic interactions between multiple threats (Gardner et al. 2009; Lees et al. 2012). Our results should also be viewed through a conservative lens, given potential historical collecting biases that we believe makes it more likely that our analysis underestimated rather than overestimated change. Many species may not have been represented in the predisturbance samples that constituted our baseline due to the difficulty of collecting small-bodied canopy species relative to understory species and the fact that collecting effort is neither temporally nor spatially constant (Burgman et al. 1995; McCarthy 1998).

Conservation Biology Volume 00, No. 0, 2014

Avian Extinctions in Amazonia

8

High

Accipiter poliogaster

Spizaetus melanoleucus Dendrexetastes rufigula* G Phoenicircus carnifex Hylophilus ochraceiceps Dacnis lineata

Micrastur semitorquatus Touit huetii Lophornis gouldii Celeus torquatus* Philydor erythrocercum Phylidor ruficaudatum Myiobius atricaudus Platyrinchus platyrhynchos Lamprospiza melanoleuca

Daptrius ater Micrastur mirandollei Deroptyus accipitrinus

≈30

POSSIBLY EXTINCT

Calliphlox amethystina Piculus chrysochloros* Berlespchia rikeri Euphonia minuta

PROBABLY E EXTINCT

Low

Accipter superciliosus Pyrrhura lepida* Hylopezus macularius Sclerurus caudacutus Sclerurus rufigularis Dendrocincla merula Deconychura longicauda* Platyrinchus saturatus Corythopis torquatus Piprites chloris*

EXTINCT

Confidence of E extinction

Penelope pileata Aburria cujubi Odontophorus gujanensis Sarcoramphus papa Harpia harpyja Spizaetus ornatus* Anodorhynchus hyacinthinus* Ara macao Guarouba guarouba* Neomorphus geoffroyi Gymnoderus foetidus Periporphyrus erythromelas

Pauxi tuberosa Psophia obscura*

>100 E

≈50 E

Number of years since last record

Figure 4. Schematic of framework illustrating how time since last record interacts with confidence of extinction (bold, globally threatened species [IUCN 2013]; asterisk, included on the regional red list). The framework can be used to determine how species can be classified locally as possibly extinct, probably extinct, or extinct (adapted from Butchart et al. [2006]). Table 2. Long-term species loss from Neotropical forest regions.

Study

Region

This study

metropolitan region of Bel´em, Par´a, Brazil Shaw et al. 2013 Sierra de Los Tuxtlas, Veracruz, Mexico Patten et al. 2010 Palenque, Chiapas, Mexico Robinson 2001 Barro Colorado Island, Panama Renjifo 1999 west slope, Cordillera Central, Colombia Ribon et al. 2003 Vic¸osa, Minas Gerais, Brazil Christiansen & Pitter 1997 Lagoa Santa, Minas Gerais, Brazil

Biome Amazonia

1812–1980, 168

47, 360, 14.5

0.28

Central America

1973–2004, 30

12, 122, 9.8

0.40

Central America 1900–2009, 109

23, 240, 9.5

0.21

Central America

1970–1996, 25

27, 200, 13.5

1.08

Andes

1911–1997, 86

6, 139, 4.3

0.06

Atlantic Forest

1932–1999, 67

28, 221, 13

0.42

Atlantic Forest

1870–1987, 117

13, 107, 12.1

0.11

Patterns of Local Extinction in the MRB Avifauna Extinction proneness in Amazonian birds is typically linked to life history characteristics such as body size, feeding behavior, and dispersal ability (Lees & Peres 2009, 2010; Stouffer et al. 2011). Our findings are con-

Conservation Biology Volume 00, No. 0, 2014

Number of extinct species, total species richness, community regionally Species extinct (%) loss/year

Period (years), no. of years

sistent with those of previous studies (e.g., Owens & Bennett 2000), which found that large-bodied species are particularly extinction prone. For example, we report the purported loss of 6 large-bodied (>1 kg) species from the MRB, of which 4—Dark-winged Trumpeter,

Moura et al.

White-crested Guan (Penelope pileata), Red-throated Piping-Guan (Aburria cujubi), and Razor-billed Curassow (Pauxi tuberosa)—are game birds highly sought after for bushmeat (e.g., Peres 2000). Eight of the largebodied species were unrecorded after 1900. The mean mass of species that went extinct from 1800 to 1900 was 1772 g, in contrast to a mean of 270 g after 1901. Given this information, the most parsimonious explanation for this first wave of local extinctions from the MRB is hunting of large-bodied species for food, although local forest loss also began to gain momentum in the same period and was most severe after 1880 (Vieira et al. 2007). Trade and persecution may also have driven some species to extinction. For example, other large-bodied species such as Scarlet Macaw, Red-and-green Macaw, and Hyacinth Macaw (Anodorhynchus hyacinthinus) were harvested for the wild bird trade in the MRB (Alves et al. 2013). The same is true of the Golden Parakeet, which we (N.G.M. and A.C.L.) found persisting in the neighboring fragmented landscapes of Moju, Paragominas, and Tailˆandia but which remains a target for the (now illegal) wild bird trade (Alves et al. 2013). Large raptors, such as Harpy Eagle (Harpia harpyja), Crested Eagle (Morphnus guianensis), and Ornate Hawk-Eagle (Spizaetus ornatus) are also particularly threatened in fragmented landscapes from hunting, which may be triggered by human–wildlife conflicts when raptors are suspected of killing small livestock (e.g., Trinca et al. 2008). Many of the other probable extinctions we found were more likely related to forest loss and disturbance than direct persecution. A wave of extinctions of smaller bodied species occurred after 1900, and the mean bodyweight of species assumed to go extinct between 1900 and 1980 was 270 g (an 85% drop in body size compared with losses observed during the previous century). This second wave of extinctions may be linked to habitat loss and degradation associated with the construction of the railway in 1883–1908 (Vieira et al. 2007) and the subsequent increase in human population and the size of settlements in Bel´em after the construction of the Bel´em–Bras´ılia road in the 1960s. The disappearance of both medium and small-bodied primary-forest dependent frugivores such as the Guianan Red-Cotinga and Wing-barred Piprites (Piprites chloris griseicens) is consistent with results of other studies (e.g., Lees & Peres 2008; Espa˜ nola et al. 2013) and may reflect loss of food resources or access to adequate nest sites. The absence of many such species was observed by early naturalists. For example, in 1926 J. Bond failed to find Red-and-Black Grosbeak (Periporphyrus erythromelas) in the MRB, remarking that it was “found only in the virgin forest at Castanhal” 50 km from Bel´em (Stone 1928). Insectivorous species have also been affected disproportionately by habitat loss, degradation, and fragmentation, including both flock-following under and mid-story primary-forest dependent species such as Long-tailed

9

Woodcreeper (Deconychura longicauda zimmeri) and terrestrial solitary species such as Spotted Antpitta (Hylopezus [macularius] paraensis), which appear intolerant even to low-intensity selective-logging activities (Lees & Peres 2010; Moura et al. 2013). The potential lack of cavity trees could have contributed to population collapses in the MRB because many of the species inferred to be locally extinct are either primary (woodpeckers) or secondary cavity nesters (e.g., forest falcons, parrots, and woodcreepers) for which habitat may be reduced in selectively logged and secondary forests (e.g., Cockle et al. 2010; Supporting Information). The Future of the MRB Avifauna Most studies documenting long-term extinctions in formerly forested tropical landscapes also report on colonization events by nonforest taxa (e.g., Patten et al. 2010). However, these marginal gains in new species do not compensate for the losses because the few colonizing species are not of conservation concern. Thus, maintenance and protection of well-preserved primary forest habitats is a prerequisite for both local (Moura et al. 2013) and global avian biodiversity conservation (Gibson et al. 2011). Future records of mobile large-bodied species may more likely represent recolonization rather than low-level persistence throughout the survey period (examples in Supporting Information). Large-bodied (often frugivorous) species with high dispersal capacity are more likely to be recorded as occasional vagrants or colonizing individuals than small-bodied (often insectivorous) species with lower dispersal capacity (Lees & Peres 2009). Although this could be grounds for guarded optimism, biotic impoverishment driven by fragmentation, unsustainable forest management, wildfires, and hunting is still ongoing in the region (Amaral et al. 2012). Given these impacts, coupled with the lack of an adequate protected area network in the MRB, we anticipate that the number of local extinctions will continue to rise and result in an ever more impoverished avifauna in this biologically unique corner of the world’s largest remaining tropical forest. Our ability to catalog the long-term erosion of biological diversity in the MRB was made possible only due to the long history of natural history research in the region. It is likely that similar, hitherto unrecorded, processes of erosion and species extinction are happening elsewhere in Amazonian deforestation frontiers and across the tropics.

Acknowledgments We thank the Instituto Nacional de Ciˆencia e Tecnologia—Biodiversidade e Uso da Terra na Amazˆ onia (CNPq 574008/2008-0), the National Environment

Conservation Biology Volume 00, No. 0, 2014

10

Research Council (NE/G000816/1), the Darwin Initiative (17-023), the Coordenac¸˜ao de Aperfeic¸oamento de Pessoal de N´ıvel Superior-CAPES, Lancaster University, Embrapa Amazonia Oriental (SEG: 02.08.06.005.00), and The Nature Conservancy for funding and The Nature Conservancy for access to detailed land-use maps of the municipality. We are very grateful to all the farmers and landowners of Paragominas for collaborating in the project, particularly Mauro L´ ucio de Castro Costa and the Sindicato dos Produtores Rurais de Paragominas. N.G.M. was supported by a CNPq doctoral studentship, A.C.L. thanks CNPq for funding, and A.A. thanks CNPq for a research productivity fellowship. We thank P. Cerqueira, M. Henriques, W. Overal, and I. Vieira for sharing data and discussions on local land-use change history. We also thank M. A. Patten and one anonymous referee whose comments helped to improve the manuscript. This paper is number 22 in the Rede Amazˆ onia Sustent´avel publication series.

Supporting Information Detailed information on the study site and data sources used to compile the sighting records (Appendix S1), additional literature (Appendix S2), threats, Red List evaluations and status of the species in the study regions (Appendix S3), and body mass of species unrecorded in the MRB after 1980 (Appendix S4) are available online. The authors are solely responsible for the content and functionality of these materials. Queries (other than absence of the material) should be directed to the corresponding author.

Literature Cited Aleixo, A. 2006. Oficina de trabalho “Discuss˜ao e elaborac¸˜ao da lista de esp´ecies ameac¸adas de extinc¸˜ao do estado do Par´a Relat´ orio T´ecnico.” Museu Paraense Em´ılio Goeldi. Available from http:// www2.museu-goeldi.br/biodiversidade/biota (accessed May 2013). Almeida, S. A., and I. C. G. Vieira. 2010. Centro de Endemismo Bel´em: Status da Vegetac¸˜ao Remanescente e Desafios para a Conservac¸˜ao Biol´ ogica e Restaurac¸˜ao Ecol´ ogica. Revista de Estudos Universit´arios 36:95–111. Alves, R. R. N., J. R. D. F. Lima, and H. F. P Ara´ ujo. 2013. The live bird trade in Brazil and its conservation implications: an overview. Bird Conservation International 23:53–65. Amaral, D. D., I. C. G. Vieira, R. P. Salom˜ao, S. S. Almeida, and M. A. G. Jardim. 2012. The status of conservation of urban forests in eastern Amazonia. Brazilian Journal of Biology 72:257–265. Bird, J. P., G. M. Buchanan, A. C. Lees, R. P. Clay, P. F. Develey, I. Y´epez, and S. H. Butchart. 2011. Integrating spatially explicit habitat projections into extinction risk assessments: a reassessment of Amazonian avifauna incorporating projected deforestation. Diversity and Distributions 18:273–281. Birdlife International. 2013. World bird database. Available from www.birdlife.org/datazone (accessed May 2013).

Conservation Biology Volume 00, No. 0, 2014

Avian Extinctions in Amazonia

Brooks, T. M., S. L. Pimm, and J. O. Oyugi. 1999. Time lag between deforestation and bird extinction in tropical forest fragments. Conservation Biology 13:1140–1150. Burgman, M., R. Grimson, and S. Ferson. 1995. Inferring threat from scientific collections. Conservation Biology 9:923–928. Butchart, S. H. M., A. J. Stattersfield, and T. M. Brooks. 2006. Going or gone: defining “Possibly Extinct” species to give a truer picture of recent extinctions. Bulletin of the British Ornithologists’ Club 126:7–24. Christiansen, M. B., and E. Pitter. 1997. Species loss in a forest bird community near Lagoa Santa in southeastern Brazil. Biological Conservation 80:23–32. CBRO (Comitˆe Brasileiro de Registros Ornitol´ ogicos). 2011. Listas das aves do Brasil. Available from www.cbro.org.br (accessed May 2013). Cockle, K. L., K. Martin, and M. C. Drever. 2010. Supply of tree-holes limits nest density of cavity-nesting birds in primary and logged subtropical Atlantic forest. Biological Conservation 143:2851– 2857. Costello, M. J., R. M. May, and N. E. Stork. 2013. Can we name Earth’s species before they go extinct? Science 339:413–416. Crowley, B. E. 2011. Extinction and rediscovery: where the wild things are. Journal of Biogeography 38:1633–1634. Espa˜ nola, C. P., N. J. Collar, and S. J. Marsden. 2013. Are populations of large-bodied avian frugivores on Luzon, Philippines, facing imminent collapse? Animal Conservation 16:467–479. Gardner, T. A., J. Barlow, R. L. Chazdon, R. Ewers, C. A. Harvey, C. A. Peres, and N. Sodhi. 2009. Prospects for tropical forest biodiversity in a human-modified world. Ecology Letters 12:561–582. Gardner, T. A., et al. 2013. A social and ecological assessment of tropical land uses at multiple scales: the Sustainable Amazon Network. Philosophical Transactions of the Royal Society B 368(1619): 20120166. Gibson, L., et al. 2011. Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478:378–381. Hansen, M. C., et al. 2013. High-resolution global maps of 21st-century forest cover change. Science 342:850–853. Hellmayr, C. E. 1905. Notes on a collection of birds, made by Mons. A. Robert in the District of Par´a, Brazil. Novitates Zoologicae 12: 269–305. INPE (Instituto Nacional de Pesquisas Espaciais). 2013. Monitoramento da floresta amazˆ onica por sat´elite: Projeto Prodes, S˜ao Jos´e dos Campos, S˜ao Paulo. Available from http://www.obt.inpe.br/prodes/ index/html (accessed August 2013). IUCN (International Union for Conservation of Nature). 2013. The IUCN red list of threatened species. Version 2013.2. Available from http://www.iucnredlist.org (accessed May 2013). Kattan, G. H., H. Alvarez-L´ opez, and M. Giraldo. 1994. Forest fragmentation and bird extinctions: San Antonio eighty years later. Conservation Biology 8:138–146. Laurance, W. F., et al. 2011. The fate of Amazonian forest fragments: a 32-year investigation. Biological Conservation 144:56–67. Le˜ao, N., C. Alencar, and A. Ver´ıssimo. 2007. Bel´em Sustent´avel 2007. Imazon, Bel´em. Lees, A. C., and C. A. Peres. 2008. Avian life-history determinants of local extinction risk in a hyper-fragmented Neotropical forest landscape. Animal Conservation 11:128–137. Lees, A. C., and C. A. Peres. 2009. Gap-crossing movements predict species occupancy in Amazonian forest fragments. Oikos 118: 280–290. Lees, A. C., and C. A. Peres. 2010. Habitat and life history determinants of antbird local extinction in variable-sized Amazonian forest fragments. Biotropica 42:614–621. Lees, A. C., N. G. Moura, A. Santana, A. Aleixo, J. Barlow, E. Berenguer, J. Ferreira, and T. A. Gardner. 2012. Paragominas: a quantitative baseline inventory of an eastern Amazonian avifauna. Revista Brasileira de Ornitologia 20:93–118.

Moura et al.

Lovejoy, T. E. 1971. Diversity and abundance patterns of lower Amazonian rain forest birds. PhD thesis, Yale University, New Haven, Connecticut. Machado, A. B. M., G. M. Drummond, and A. P. Paglia. 2008. Livro vermelho da fauna brasileira ameac¸ada de extinc¸˜ao. 1st edition. MMA; Belo Horizonte, MG: Fundac¸˜ao Biodiversitas, Bras´ılia, DF. McCarthy, M. A. 1998. Identifying declining and threatened species with museum data. Biological Conservation 83:9–17. McInerny, G. J., D. L. Roberts, A. J. Davy, and P. J. Cribb. 2006. Significance of sighting rate in inferring extinction and threat. Conservation Biology 20:562–567. Metzger, J. P., A. C. Martensen, M. Dixo, L. C. Bernacci, M. C. Ribeiro, A. M. G. Teixeira, and R. Pardini. 2009. Time-lag in biological responses to landscape changes in a highly dynamic Atlantic forest region. Biological Conservation 142:1166–1177. Moura, N. G., A. C. Lees, C. B. Andretti, B. J. W. Davis, R. R. C. Solar, A. Aleixo, J. Barlow, J. Ferreira, and T. A. Gardner. 2013. Avian biodiversity in multiple-use landscapes of the Brazilian Amazon. Biological Conservation 167:339–348. Novaes, F. C. 1973. Aves de uma vegetac¸˜ao secund´aria na foz do Amazonas. Publicac¸˜ao Avulsa Museu Paraense Emilio Goeldi 21:1–88. Novaes, F. C., and M. D. F. C. Lima. 2009. Aves da grande Bel´em: munic´ıpios de Bel´em e Ananindeua, Par´a. PR/McT/Cnpq. Owens, I. P., and P. M. Bennett. 2000. Ecological basis of extinction risk in birds: habitat loss versus human persecution and introduced predators. Proceedings of the National Academy of Sciences of the United States of America 9:12144–12148. Patten, M. A., H. G. Silva, and B. D. Smith-Patten. 2010. Long-term changes in the bird community of Palenque, Chiapas, in response to rainforest loss. Biodiversity and Conservation 19:21–36. Peres, C. A. 2000. Effects of subsistence hunting on vertebrate community structure in Amazonian forests. Conservation Biology 14: 240–253. Peres, C. A., T. A. Gardner, J. Barlow, J. Zuanon, F. Michalski, A. C. Lees, I. C. G. Vieira, F. M. S. Moreira, and K. J. Feeley. 2010. Biodiversity conservation in human-modified Amazonian forest landscapes. Biological Conservation 143:2314–2327. Pickett, S. T. 1989. Space-for-time substitution as an alternative to longterm studies. Pages 110–135 in G. E. Likens, editor. Long-term studies in ecology: approaches and alternatives. Springer-Verlag, New York. Portes, C. E. B., L. S. Carneiro, F. Schunck, M. S. S. Silva, K. J. Zimmer, A. Whittaker, F. Poletto, L. F. Silveira, and A. Aleixo. 2011. Annotated checklist of birds recorded between 1998 and 2009 at nine areas in

11

the Bel´em area of endemism, with notes on some range extensions and the conservation status of endangered species. Revista Brasileira de Ornitologia 19:167–184. Renjifo, L. M. 1999. Composition changes in a subandean avifauna after long-term forest fragmentation. Conservation Biology 13:1124– 1139. Ribon, R., J. E. Simon, and G. T. Mattos. 2003. Bird extinctions in Atlantic forest fragments of the Vic¸osa region, southeastern Brazil. Conservation Biology 17:1827–1839. Robinson, W. D. 2001. Changes in abundance of birds in a Neotropical forest fragment over 25 years: a review. Animal Biodiversity and Conservation 24:51–65. Sclater, P. L., and O. Salvin. 1867. List of birds collected by Mr. Wallace on the lower Amazons and Rio Negro. Proceedings of the Zoological Society of London 1867:566–596. Shaw, D. W., P. Escalante, J. H. Rappole, M. A. Ramos, R. J. Oehlenschlager, D. W. Warner, and K. Winker. 2013. Decadal changes and delayed avian species losses due to deforestation in the northern Neotropics. PeerJ 1. DOI:10.7717/peerj.179. Sodhi, N. S., L. H. Liow, and F. A. Bazzaz. 2004. Avian extinctions from tropical and subtropical forests. Annual Review of Ecology and Systematics 35:323–345. Solow, A. R. 1993. Inferring extinction from sighting data. Ecology 74:962–964. Solow, A. R. 2005. Inferring extinction from a sighting record. Mathematical Biosciences 195:47–55. Stone, W. 1928. On a collection of birds from the Par´a region, eastern Brazil (with field notes by J. Bond and R. Meyer de Schauensee). Proceedings of the Academy of Natural Sciences of Philadelphia 80:149–176. Stouffer, P. C., E. I. Johnson, R. O. Bierregaard Jr., and T. E. Lovejoy. 2011. Understory bird communities in Amazonian rainforest fragments: species turnover through 25 years post-isolation in recovering landscapes. PLoS ONE 6:e20543. Teixeira, D. M., N. Papavero, and L. B. Kury. 2010. As aves do Par´a ´ lvares Roxo de Potflis segundo as “mem´ orias” de Dom Lourenc¸o A (1752). Arquivos de Zoologia 41:97–131. Trinca, C. T., S. F. Ferrari, and A. C. Lees. 2008. Curiosity killed the bird: arbitrary hunting of Harpy Eagles (Harpia harpyja) on an agricultural frontier in southern Brazilian Amazonia. Cotinga 30: 12–15. Vieira, I. C. G., P. M. Toledo, and A. Almeida. 2007. An´alise das ˜ es da Paisagem da Regi˜ao Bragantina, no Par´a. Integrando modificac¸o diferentes escalas de tempo. Ciˆencia e Cultura 59:27–30.

Conservation Biology Volume 00, No. 0, 2014