Data acquisition

To acquire information about the winter distribution of Palearctic migratory birds in Africa, we contacted hundreds of individuals or organizations via mail, email or telephone requesting data or references. We also contacted all ringing schemes within Europe and Africa that we knew of. Over 100 people and almost all ringing schemes responded and sent data, references or contact addresses (see Acknowledgments for details).

We entered 24396 distributional records into our database so far. The bold entries below are data entered into the database, and the bold italics entries are data that can be interactively searched on the search page. Distributional data were entered as follows:

1) The species name was entered as a Latin binomial. The Birds of the Western Palearctic (Snow and Perrins 1998) was used for all Latin names, except for those names that were recently changed by Sangster et al. (2002) and Knox et al. (2002). 343 species were identified as Western Palearctic - African migrants (of which 240 species migrate to the Sub-Saharan region) after a careful study of the literature (see List of Western Palearctic bird species migrating within Africa).

2) The location of each record was entered as country (see List of countries), location, and latitude/longitude. If the source did not provide coordinates, we consulted the Times Atlas (Bartholomew 1956, Anonymous 2001), various other printed gazetteers, or the internet-based gazetteer of the National Geospatial-Intelligence Agency (2005). If these gazetteers provided better coordinates, we corrected the coordinates provided by the original sources. We entered an additional code indicating the accuracy of the location given the information in the source. The most reasonable estimate of the accuracy of the location was entered using the following codes:

0 = exact locality, i.e. whenever geographical coordinates were provided by the source except when the source suggested that one of the codes below should be used.
1 = accuracy is ± 1 minute; this code was used whenever an exact location was given in the source, e.g. the name of a town, but no coordinates were provided; this code was also used whenever the source allowed calculation of geographical coordinates, e.g. by stating the location is "between ... and" or "10 kilometres south of".
2 = accuracy is ± 5 kilometres; this code was used by a few sources.
3 = accuracy is ± 10 kilometres; this code was used by a few sources.
4 = accuracy is ± 10 minutes; this code was used whenever the source suggested that the record was somewhat inaccurate, e.g. by stating the location was within an "area", "delta", "district", "plateau", "region" or when a qualifier was attached to the location's name such as "east of" or "near".
5 = accuracy is ± 15 minutes; this code was used mainly by atlas projects.
6 = accuracy is ± 30 minutes; this code was used mainly by atlas projects.
7 = accuracy is ± 1 degree; this code was used whenever the source suggested that the record was quite inaccurate, e.g. by stating the location was within an "province".
8 = accuracy is ± 10 degrees; this code was used whenever the source suggested that the record was very inaccurate, e.g. by stating the location was within a "country".
9 = this code was used whenever coordinates were provided but their accuracy could not be ascertained.
10 = this code was used when the location unknown.

3) The date of each record was entered as day, month and year. We entered an additional code indicating the accuracy of the date given the information in the source. The most reasonable estimate of the accuracy of the date was entered as an interval of ± the number of days that the date may be inaccurate. Two unique codes were used to indicate if the uncertainty interval was ± several years (code -9) or if the date was not given or unreliable (code -99).

4) Reliability was categorised as 1) specimen obtained, 2) ringing record, 3) published observational record, or 4) unpublished observational record, received as personal communication.

5) The collector of each record was entered, which refers to the person or organization (e.g. ringing schemes) which provided the data.

6) The database number is a unique identifier number for each record in the database.

References

Anonymous 2001. The Times comprehensive atlas of the world. London, UK: Times Book Group Ltd.
Bartholomew J (ed). 1956. The Times atlas of the world. Volume IV: Southern Europe and Africa. The Times Publishing Company Ltd, London.
Knox AG, Collinson M, Helbig AJ, Parkin DT, Sangster G. 2002. Taxonomic recommendations for British birds. Ibis 144: 707-710.
National Geospatial-Intelligence Agency 2005. GeoNet Names Server (GNS). http://gnswww.nga.mil/geonames/GNS/index.jsp
Sangster G, Knox AG, Helbig AJ, Parkin DT. 2002. Taxonomic recommendations for European birds. Ibis 144: 153-159.
Snow DW, Perrins CM. 1998. The birds of the western Palearctic: concise edition. Vols. 1-2. Oxford University Press, Oxford, UK.