The Taiwan Breeding Bird Survey Data

Study Extent

Sampling sites of BBS Taiwan were selected to cover all major habitats on the main island of Taiwan and also the surrounding islands. Currently, the dataset contains 480 sampling sites, which includes a total of 4,138 survey points. The major habitat types covered by these sampling sites include forest (48%), farmland (24%), urban (12%), wetlands (7%) and grasslands (6%).

Sampling

BBS Taiwan adopts the point-count method (Bibby et al. 2000) with a 6 minutes count duration per point.

Quality Control

Efforts to maintain the quality of data were implemented before and after the survey. Before the survey, training workshops with field excursions were held to ensure participants fully understand the survey methods. After the data were submitted by the surveyors, location of survey points, correctness of the time period, distance, flock notes were reviewed. Filters for rare or easily-misidentified species and unusual high counts of numbers were also used to detect possibly error records. The alerted records were then double-checked with surveyors, and doubtful records were deleted from the release version of the dataset. The taxonomic system follows the Checklist of Birds of Taiwan maintained by the Bird Record Committee of the Chinese Wild Bird Federation (Ding et al. 2017). The dataset described and provided here is essentially the complete set of the collected raw data.

Method steps

1. BBS Taiwan applied a stratified random sampling scheme for an initial set of 450 sampling sites aimed at covering 5% of the c.f. 36,000 km2 area of the Taiwan Island. Nighty-one stratums, defined by 41 eco-regions (Su 1992) combined with three different elevation zones (0-1,000m, 1,000-2,500m, 2,500-4,000m) where used in this progress. An additional “customized” set of samplings sites were selected manually by surveyors at locations where no pre-selected sites covers. To date, all types of sampling sites now reach approximately a number of 750 sites. All sampling sites were available to be adopted by volunteered surveyors each year, while encouraging the same surveyor to adopt the same sampling site as long as possible. Each sampling sites were surveyed twice a year, which, in combination with the 6 – 10 survey points per site, gives each site a minimum sampling effort of 12 times. The above sampling effort ensures the likelihood of detecting 80% of species present and keeping the coefficient of variance below 20 % in the breeding season within each site on Taiwan (Ding 1993, Shiu and Lee 2003). Surveys are performed within the breeding season between March and June, with sites at higher elevations surveyed in a later sequence. Surveys are conducted within four hours of sunrise on each day. Point-count method (Bibby et al. 2000) with a 6-minute stop time were conducted on each survey point, ensuring the effort to detect a minimum of 80% of the species at each site (Ding 1993, Shiu and Lee 2003). We record every bird individuals detected, regardless of migratory status and either sighted or heard. For each bird record, the period of detection time (0-3 minutes, 3-6 minutes, supplementary which are records beyond the 6-minute stop time), horizontal distance to the observer (0-25m, 25-100m, >100m., flyover), and formation of flock with more than five individuals were recorded. For each survey point, the main-habitat, sub-habitat, presence of water-bodies, weather condition, starting time were recorded. Habitat were classified by surveyors on field.

A total of 313 bird species from 62 families are recorded, including all 27 endemic bird species on the island, and also introduced bird species.

1. Ave
   common name: Bird rank: class

The records in the BBS Taiwan dataset are located on the main island of Taiwan and also the Lanyu Island off the east coast. Environmental gradients of the sampling locations spans over the elevation zone from 0-3,900m ASL, and across habitats from human-dominated urban landscapes to primary forests in the remote mountain areas. Each sampling site was composed of 6-10 survey points with a 100 meter radius. Coverage of sampling sites is expected to increase in the future.