Because of its larger size, we visited Bintan 10 times (21–23 November 2006, 1–7 February, 1–7 April, 27–30 May, 18–24 June, 24–27 August, 24–26 September, 16–22 October, 1–7 December 2007, and 11–12 March 2008). During the first two visits, the whole island was surveyed to locate areas containing secondary and primary forests. Each of these forested sites was visited between four and 12 times. All surveys were made during fair weather (no heavy rain or wind) and between 06:00 and 11:00 and 15:00–18:00. At randomly selected locations in forests,we also played the songs or calls of 15 missing species for 3 min each, followed by a 5-min period in which we
waited for the species to respond. These playback surveys were conducted at all eight forested sites containing late secondary (>50 years old) or old-growth forests. Once a species’ presence was reconfirmed, its call was no longer played.We did not conduct song or call playback surveys
on other islands because of lack of appropriate recordings and time.
Some of the surveyed islands were also sampled between 1977 and 1996, which provided additional
data to validate extinctions (Wells 1977; Sheldon et al. 1983; Rajathurai 1996). For regression analysis, we also used published data on bird extinctions from Singapore (Castelletta et al. 2000; Brook et al. 2003; Table 1).
The completeness of our sampling efforts across all islands was assessed by constructing sample-based rarefaction curves derived from incidence data (EstimateS software, version 8.0; Colwell 2006). In addition, we estimated species richness with eight nonparametric richness estimators (abundance-based coverage, incidencebased coverage, Chao 1 and 2, Jack 1 and 2, bootstrap,
and MMRuns; Colwell 2006) to assess sampling adequacy. The higher the association between observed and estimated richness, the more complete our sampling efforts are. We classified a species as locally extinct (extirpation) if it was not found in the current (2006–2008) or more
recent (1977–1996) surveys. Following MacKinnon and Phillips (1993), we considered those forest species that preferred forested habitats or were found mostly in oldgrowth forests (e.g., Scarlet-rumped Trogon [Harpactes duvaucelii]). Results of our species inventories across
the islands are provided in Supporting Information. We excluded records of bird species that had not been observed previously on the islands in our analyses because it was unclear whether these were new colonizations or species were missed by previous surveyors.
Because of its larger size, we visited Bintan 10 times (21–23 November 2006, 1–7 February, 1–7 April, 27–30 May, 18–24 June, 24–27 August, 24–26 September, 16–22 October, 1–7 December 2007, and 11–12 March 2008). During the first two visits, the whole island was surveyed to locate areas containing secondary and primary forests. Each of these forested sites was visited between four and 12 times. All surveys were made during fair weather (no heavy rain or wind) and between 06:00 and 11:00 and 15:00–18:00. At randomly selected locations in forests,we also played the songs or calls of 15 missing species for 3 min each, followed by a 5-min period in which wewaited for the species to respond. These playback surveys were conducted at all eight forested sites containing late secondary (>50 years old) or old-growth forests. Once a species’ presence was reconfirmed, its call was no longer played.We did not conduct song or call playback surveyson other islands because of lack of appropriate recordings and time.Some of the surveyed islands were also sampled between 1977 and 1996, which provided additionaldata to validate extinctions (Wells 1977; Sheldon et al. 1983; Rajathurai 1996). For regression analysis, we also used published data on bird extinctions from Singapore (Castelletta et al. 2000; Brook et al. 2003; Table 1).The completeness of our sampling efforts across all islands was assessed by constructing sample-based rarefaction curves derived from incidence data (EstimateS software, version 8.0; Colwell 2006). In addition, we estimated species richness with eight nonparametric richness estimators (abundance-based coverage, incidencebased coverage, Chao 1 and 2, Jack 1 and 2, bootstrap,and MMRuns; Colwell 2006) to assess sampling adequacy. The higher the association between observed and estimated richness, the more complete our sampling efforts are. We classified a species as locally extinct (extirpation) if it was not found in the current (2006–2008) or morerecent (1977–1996) surveys. Following MacKinnon and Phillips (1993), we considered those forest species that preferred forested habitats or were found mostly in oldgrowth forests (e.g., Scarlet-rumped Trogon [Harpactes duvaucelii]). Results of our species inventories acrossthe islands are provided in Supporting Information. We excluded records of bird species that had not been observed previously on the islands in our analyses because it was unclear whether these were new colonizations or species were missed by previous surveyors.
การแปล กรุณารอสักครู่..
Because of its larger size, we visited Bintan 10 times (21–23 November 2006, 1–7 February, 1–7 April, 27–30 May, 18–24 June, 24–27 August, 24–26 September, 16–22 October, 1–7 December 2007, and 11–12 March 2008). During the first two visits, the whole island was surveyed to locate areas containing secondary and primary forests. Each of these forested sites was visited between four and 12 times. All surveys were made during fair weather (no heavy rain or wind) and between 06:00 and 11:00 and 15:00–18:00. At randomly selected locations in forests,we also played the songs or calls of 15 missing species for 3 min each, followed by a 5-min period in which we
waited for the species to respond. These playback surveys were conducted at all eight forested sites containing late secondary (>50 years old) or old-growth forests. Once a species’ presence was reconfirmed, its call was no longer played.We did not conduct song or call playback surveys
on other islands because of lack of appropriate recordings and time.
Some of the surveyed islands were also sampled between 1977 and 1996, which provided additional
data to validate extinctions (Wells 1977; Sheldon et al. 1983; Rajathurai 1996). For regression analysis, we also used published data on bird extinctions from Singapore (Castelletta et al. 2000; Brook et al. 2003; Table 1).
The completeness of our sampling efforts across all islands was assessed by constructing sample-based rarefaction curves derived from incidence data (EstimateS software, version 8.0; Colwell 2006). In addition, we estimated species richness with eight nonparametric richness estimators (abundance-based coverage, incidencebased coverage, Chao 1 and 2, Jack 1 and 2, bootstrap,
and MMRuns; Colwell 2006) to assess sampling adequacy. The higher the association between observed and estimated richness, the more complete our sampling efforts are. We classified a species as locally extinct (extirpation) if it was not found in the current (2006–2008) or more
recent (1977–1996) surveys. Following MacKinnon and Phillips (1993), we considered those forest species that preferred forested habitats or were found mostly in oldgrowth forests (e.g., Scarlet-rumped Trogon [Harpactes duvaucelii]). Results of our species inventories across
the islands are provided in Supporting Information. We excluded records of bird species that had not been observed previously on the islands in our analyses because it was unclear whether these were new colonizations or species were missed by previous surveyors.
การแปล กรุณารอสักครู่..