This dataset includes observations of Macaroni Penguin (Eudyptes chrysolophus) that are sourced from the Atlas of Living Australia (ALA) database. Rather than raw observations, these have been filtered such that they are assumed to be suitable for species distribution modelling exercises. The cleaning process included:

1. automatic vetting based on the ALA's 'assertions' whereby observations were assessed as inappropriate for modelling (ie. 'ZERO_COORDINATES', 'INVALID SCIENTIFIC NAME');
2. determining if the observations fell within expert-derived range polygons. These polygons were supplied by BirdLife Australia to represent, for each species, its core breeding habitat, non-breeding, historic, irruptive, or invasive ranges. Records that fall outside these ranges were marked as inappropriate for modelling; and
3. human-derived classification of records after previous two assessments. Through the Edgar project (http://tropicaldatahub.org/goto/edgar), users were able to map all species observations and comment on the suitability of records for distribution modelling. This included records deemed inappropriate by other means.

   Every 6 months the occurrence record download file is updated to reflect recent vetting by experts. In the data download, sensitive records have been obfuscated by truncating the lat/long to two decimal places. Obfuscated records will be indicated in the data file. Access to the accurate data will need to be arranged with the original data owners - contact the ALA for more information.

   The resulting downloadable file of occurrence records reflects which records are suitable for species distribution modelling.

Data is downloadable as a zipped CSV file.

These occurrence records are displayed on Edgar: http://tropicaldatahub.org/goto/Edgar.