Study area description
Valid distribution records are located in the northern portion of the Neotropical region, including the transitional zone with the Nearctic region. Native vegetation in this area ranges from tropical dry to humid forests, and from lowlands to montane forests. However, a large proportion of the native vegetation has been converted to pasture and agricultural areas. The expansion of human settlements, infrastructure, and mining have also contributed to forest degradation and deforestation in the region. Tropical forests have the largest net loss of forested area compared to other forest types in the world (FAO and JRC 2012), and the Neotropical region is not the exception. The study area includes the Mesoamerica biodiversity hotspot, the Chocó/Darién/Western Ecuador hotspot, and marginally the Tropical Andes hotspot (Myers et al. 2000), but these hotspots harbour only 20 to 25% of the original extent of primary vegetation.
Given the current pattern of forest cover in the region, and the temporal coverage of records in the CracidMex1 database, many records, particularly older records, are now located outside of currently forested areas. This suggests a substantial reduction in the distribution of cracid species, particularly for those species restricted to primary forests (P. purpurascens, P. nigra, O. derbianus, and C. rubra). Habitat loss and hunting pressure are the main drivers of cracid population declines and distribution contractions, the synergy of which has caused the endangerment of these species (Silva and Strahl 1991, 1997; Brooks and Strahl 2000; del Hoyo and Motis 2004).
The construction of the CracidMex1 database aimed to gather most of the globally available records of cracids which are distributed in Mexico, in order to generate global species distribution models. We initiated the construction of the database by collating records from six electronic databases available through the Internet: GBIF <http://data.gbif.org>, ORNIS <http://www.ornisnet.org>, REMIB <http://www.conabio.gob.mx/remib/doctos/remib_esp.html>, UNIBIO <http://unibio.unam.mx>, SpeciesLink <http://splink.cria.org.br>, and IBC <http://ibc.lynxeds.com>. Additionally, we obtained records from the National System of Information on Biodiversity (SNIB) database at CONABIO and from museum specimen records contained in the Bird Atlas of Mexico database at the Facultad de Ciencias of the National Autonomous University of Mexico. We also obtained records from published papers through searches in BioOne <http://www.bioone.org>, EBSCO <http://search.ebscohost.com>, JSTOR <http://www.jstor.org>, ScienceDirect <http://www.sciencedirect.com>, Springer Link <http://www.springerlink.com>, Web of Science <http://apps.webofknowledge.com>, Wiley Online Library <http://onlinelibrary.wiley.com>, Zoological Record <http://thomsonreuters.com/zoological-record/>, Redalyc <http://www.redalyc.org>, SciELO <http://www.scielo.org>, and Google Scholar . We also reviewed the bulletins of the Cracid Group of the Galliformes Specialists Group <http://www.cracids.org>. Added to which, we gathered records from “grey literature” through searches in technical reports and theses. These searches included the electronic portal of CONABIO and the repositories OpenDOAR <http://opendoar.org> and the Registry of Open Access Repositories <http://roar.eprints.org>. Finally, we gathered records from our own and unpublished databases of colleagues through personal contacts. After the GBIF, these personal unpublished databases were the second most important source of records, followed by records gathered from the SNIB and published papers (Table 4).
Database quality control, based on the standards described in CONABIO (2012), was an iterative process that commenced with the detection, consolidation and elimination of duplicate records (the same record reported in more than one source). For detection of duplicate records within and among sources we first gave priority to the fields “institutionCode”, “catalogNumber”, “country”, “state”, “locality”, “decimalLatitude”, and “decimalLongitude”. The consolidation process consisted of the creation of a single record with more complete data from duplicate records. In the case of inconsistencies in duplicate records, we referred to the original source of the record. We avoided and corrected errors (omission, typographic, contextual, redundancy, convention, and congruence) through automatized tasks and case by case revision of the database. We then calculated geographic coordinates and their uncertainties for those records lacking these data, based on the standards described in CONABIO (2008). All coordinates refer to the datum WGS84. We used a variety of resources for geo-referencing, namely Google Earth 7 <http://www.google.com/earth/index.html>, Google Maps and the tools of Map Labs <http://maps.google.com>, glosk <http://www.glosk.com/>, CONABIO <http://www.conabio.gob.mx/informacion/metadata/gis/loc2000gw.xml?_httpcache=yes&_xsl=/db/metadata/xsl/fgdc_html.xsl&_indent=no>, GEOSiB <http://www.humboldt.org.co/geoinformacion/geosib>, and Georeferencing Calculator <http://manisnet.org/gci2.html>. We also consulted regional experts for advice during the geo-referencing process. Once we were sufficiently certain of the correct location of the record, we checked that each location was consistent with taxa identification by displaying the records in a GIS. This taxonomic and geographic validation through the use of GIS tools and expert knowledge allowed us to detect inconsistencies. Where possible, we corrected inconsistencies through an iterative process, otherwise we labelled the record as “doubtful” (979 records) or “absent” (186) in the “occurrenceStatus” field as described above.
The CracidMex1 database has 41 fields based on the standard Darwin Core version 1.4.
National Commission for Knowledge and Use of Biodiversity (CONABIO), Mexico, under the agreement FB1585/JM024/12.