email a friend
printable version
Terms & definitions

Click on any of the following:

Species populations and trends

In the species factsheets, population and trend information are given where available. In BirdLife's World Bird Database, population and trend data are held such that they can be tested against the IUCN Red List criteria. Some of these fields are listed in the factsheet additional data tables, including the number of mature individuals, population trend, number of subpopulations, size of the largest subpopulation and generation length (see IUCN Red List Criteria for definitions). Where specific population sizes have not been estimated, estimates have been made in one of the standard bands relevant to the IUCN Red List criteria, i.e. 1-49, 50-249, 250-999, 1,000-2,499, or 2,500-10,000 individuals.

Population justification - This field gives the source of the population estimate where this has been published. For a number of species with unknown population sizes, population estimates have been derived from expert judgements by specialists in the BirdLife network. To check the validity of these we have calculated potential population size ranges using values derived from the (a) estimated Area of Occupancy, and (b) published or unpublished population density estimates for related species.

Area of Occupancy - For most Globally Threatened Birds (GTBs), Extent of Occurrence (EOO) has been calculated from maps based on known and projected range. However, only part of the EOO is occupied as the Area of Occupancy (AOO). An analysis of Globally Threatened Birds and Australian Action Plan species (Garnett and Crowley 2000) with independent estimates for both values was carried out. For both datasets, seabirds and waterbirds were excluded, as their habitats (seabird nesting colonies and wetlands) are fragmented and distributed in ways that are fundamentally different from the pattern of remaining habitat patches for most other birds. The analysis showed that AOO as a proportion of EOO increases as EOO decreases. Although there is considerable variation, very approximate estimates of AOO can be derived by taking 10% of EOO if EOO is >10,000 km², 20% if EOO is 1,000-10,000 km², and 45% if EOO is 100-1,000 km². These percentages were therefore used to calculate a potential AOO. In some cases the percentage was increased if the species' habitat is reasonably intact and homogenous, or decreased if the species is a habitat specialist.

Population density - To date, 2,786 density estimates (c.1,500 species) for landbirds mainly in Central and South America (1,080), Africa (1,052), Asia (429), and the Pacific (199) have been collated from a total of 112 published and unpublished sources into the 'BirdLife Population Density Spreadsheet'. For each GTB assessed using the method described above, a range of population estimates was obtained from values for related species occurring in similar habitats and altitudes. For uncommon species, or those with no subjective estimate of abundance, a range of densities was taken from the lowest to the lower quartile of density values (ignoring outliers) for similar-sized related species, in similar habitat, at similar altitudes, with similar ecology. For species described as common, the density range was taken from the median to the upper quartile of density estimates. In each case the density range was then multiplied by the estimated AOO. The resulting estimates were then placed in one of the standard bands relevant to the IUCN Red List criteria (see above).

Caveats - There are many limitations with such extrapolations, including:

  1. The techniques used to estimate densities in different studies vary considerably. In the past, line transects were popular. Estimates based on these may be less reliable than Variable Circular Plot techniques which have been more popular in recent studies. However, these too make a number of untested or invalid assumptions and may produce estimates of questionable validity.
  2. Many species vary in their abundance at different sites, even those at similar altitudes and with indistinguishable habitat. Hence a density estimate at one site may not be representative of the species elsewhere.
  3. Published estimates may be biased towards commoner species, as estimates are often not made for species recorded rarely during surveys and hence those for which there is little data.
  4. Some published densities will be for species in sub-optimum habitat, and these will be lower than is typical for the species.
  5. Threatened species may be rare species which, as a result of their ecology, occur at lower densities than their non-threatened congeners. Taking density estimates from congeners would therefore lead to inflated population size estimates.
  6. Threatened species may occur at different altitudes and different habitats (and therefore at different densities) to their congeners for which we have density estimates.

However, it is hoped that being explicit about the assumptions upon which population estimates are based will stimulate greater fieldwork to provide more reliable direct population size estimates.

Generation length - Using the IUCN Red List Criterion "A", trends are considered over periods of (usually) ten years or three generation lengths (defined as the mean age of the parents of the current cohort), whichever is longer. Generation length has been estimated for all seabirds using the formula: generation length = (1/m)+b where m = mean annual mortality and b = age at first breeding. Data for these parameters was taken from the literature, in particular Schreiber and Burger (2002). For species for which either or both of these parameters were unavailable, values were extrapolated from the mean for congeners. For non-seabirds, values were taken from the literature when available, but in the vast majority of cases the default period of ten years was used. For many tropical species three generations is likely to be in excess of ten years. BirdLife is currently attempting to collate appropriate life history data for non-seabirds to allow generation length to be estimated in future.