The taxonomic coverage (Figure 2) of this database spans the full range of ants pertaining to Belgium (indigenous ant species and exotic introduced species). The determination level is species level and, if appropriate, hybrid level. For some species, information on micro-and macrogynes is available. Key milestones of FORMIDABEL from conception till date are described in the ”Dataset” section of this manuscript.

The taxonomic authorities followed are: Radchenko and Elmes (2010) for the genus Myrmica and Seifert (2007) for the other genera. The dataset contains occurrences of 76 native and 9 introduced species.

As depicted in Figure 2, the most abundant subfamily in the database is the Formicinae (49, 4%) followed by the Myrmicinae (48, 7%), the Dolichoderinae (0, 9%) and the Ponerinae (0, 8%). The top five most recorded species are Lasius niger (2846 records), Myrmica rubra (2601 records), Myrmica scabrinodis (1626 records), Formica fusca (1467 records) and Myrmica sabuleti (1202 records).

Belgium is a small country in Western Europe. To the west, its 70 km coastline fronts the North Sea; to the north lies the Netherlands; to the east, Germany, and to the south, France and Luxembourg. Biologeographically, the fauna of eastern Belgium belongs to the Central European Province of the Eurasian (Palaearctic) region. By contrast, the rest of the country primarily consists of an Atlantic fauna plus a few Central European relict species.

Politically and geographically, the country is divided into three parts: Flanders, Wallonia and the Brussels Capital Region (Figure 3). In Flanders (13, 522 km² and population about 6 million people), to the north, soils are mainly sandy to loamy. Here, the most important habitats for ants are heathlands and dry grasslands. The Brussels Capital Region is a small region (162 km²) entirely situated in the sandy loam area. In Wallonia (17, 006 km² and about 3, 5 million people), to the south, soils and habitats are more diverse, ranging from forests to rocky and calcareous grasslands on loam and chalky soils. Eastern Wallonia, near the German border, includes the Hautes Fagnes, a large area of bogs and peat with some typical ant species.

The Universal Transverse Mercator Projection (UTM), an adaptation of the standard Mercator projection, uses a two dimensional Cartesian co-ordinate system to identify locations on the surface of the Earth (Wikipedia).

The UTM 5 Km (Universal Transverse Mercator Projection) raster projection divides Belgium in approximately 1200 25 km² squares (Figure 3). A representative number of UTM squares has been sampled (1125 UTM 5×5 km squares of which 659 squares with more than 10 records: see Figures 3 and 4) to complete the dataset. All the records in FORMIDABEL are georeferenced through the centroid coordinates of the corresponding UTM 5 km square. Therefore, the uncertainty on these coordinates is 3.500 meters, the distance between the centre and the corner of the UTM square.

More than half of the records are provided with a description of the microhabitat of the record locality. This allows us to give details on ecological preferences of all Belgian ant species. In FORMIDABEL we created a list of potential microhabitats for ants in Belgium. For each of these microhabitats we use a code called the “ecocode”. This code thus gives information on the habitat were the occurrence was made. Nine types of habitat and landscape were defined to collect accurate information on the habitat preference of all ant species (Dekoninck et al. 2005). These nine types are based on the EIS-code and the Flemish nature types (Vandenbussche 2002; Zwaenepoel et al. 2002). When no habitat description was available for a record (for example, with some older records) the habitat was coded as ‘Not known or not observed’. The nine main habitats we defined were: i) anthropogenic habitats, ii) dry grasslands, iii) moist grasslands, iv) forests, v) chalk grasslands, stony slopes and other rocky xerothermic habitats, vi) shrubs, vii) heathlands, viii) fens and highland bogs and ix) coastal and inland dunes.

49°27'0"N and 51°32'24"N Latitude; 2°28'12"E and 6°27'36"E Longitude

A large portion of the occurrence data have been collected by volunteers, other records originated from several projects and research programs. The data and specimens were sent to the Belgian ant curators, and after validation, the information was incorporated in the database. The collection records “dry specimen” originate from the Gembloux “Ant” collection and the Charles Gaspar collection, the collection of the “Cercle des entomologists Liégeois”, the RBINS collection and the private collection “Roland Vannieuwenhuyse”. After revision and validation, this information was also included in the database. The literature-based records were retrieved from van Boven 1970; van Boven and Mabelis 1986; Dekoninck et al. 2006 and references therein. How the database evolved is described in the Database history section.

Sampling description: Most occurrence records originate from hand/nest sampling (42, 3% of all records and mainly from Wallonia) andpitfall sampling (36, 7% mainly from Flanders). The followed procedure differs from region to region. This is due to historical reasons. Some very interesting occurrence records were obtained by sifting, coloured water traps and Malaise traps (all less than 3 % of the total sampling). An extensive description of the sampling methods is provided by Schauff (2001).

Quality control description: All the records were validated by the dataset curators before being added to the FORMIDABEL database. The dataset curators also checked the determinations of the collection specimens. If needed, the determination was adapted and made consistent with modern taxonomy; Radchenko and Elmes (2010) for the genus Myrmica and Seifert (2007) for the other genera. Before the final publication of the database all the records were tested on geographical consistency by the Belgian Biodiversity Platform and corrected if necessary.

At the beginning of 2001 all available records of ants in Flanders (northern part of Belgium) were brought together for the first time and several inventories were started. More than 20.000 records (for the most part gathered after 1990) were assembled in the FORMIDABEL (FORMIcidaeDAtaBELgium) database resulting in the ‘Verspreidingsatlas en voorlopige Rode lijst van de mieren van Vlaanderen; Dekoninck et al. 2003. [Distribution atlas and preliminary Red list of ant species in Flanders, Belgium]. This report contains all available distribution data for Flanders. In the southern part of Belgium (Wallonia) intensive inventorying started in 2005. Until then knowledge on the distribution of ants in Wallonia was limited. Only a few areas (Haute Fagnes, Famenne and the Viroin valley) had already been inventoried. Thanks to the good cooperation between the Polyergus and the FourmisWalBru working groups; many Belgian ant records were brought together in the FORMIDABEL database. Since then, FORMIDABEL is updated with data originating from FourmisWalBru twice a year. The FORMIDABEL dataset was then used for the creation of the “Belgian Ant Atlas”, (Dekoninck et al. 2012). In 2010 a cooperation agreement between the authors of the atlas and the Belgian Biodiversity Platform (www.biodiversity.be) was made. Together with the publication of the book, the data was published in Darwin Core Archives to GBIF (url: http://data.gbif.org/datasets/resource/14697/) and a dedicated data portal was created: www.formicidae-atlas.be.

The original FORMIDABEL database was created in Microsoft Access. Later, this database was completely imported in a relational SQL database. During this process additional data cleaning was performed; see section Quality control description. The dataflow is illustrated in Figure 6.

The FORMIDABEL Darwin Core Archive is a custom made SQL view on the original version of the FORMIDABEL access database which is in the custody of the Belgian Ant working groups Polyergus and fourmisWalBru. Mind that every record in FORMIDABEL represent at least one occurrence, but primarily contains information on the presence of a species. The view only shows data that are accepted for publication. Fields given are:

id, decimalLatitude, family, basisOfRecord, stateProvince, identifiedBy, eventDate, modified, country, individualCount, scientificName, kingdom, order, geodeticDatum, genus, collectionCode, decimalLongitude, samplingProtocol, catalogNumber, phylum, recordNumber, countryCode, coordinatePrecision, language, coordinateUncertaintyInMeters, locality, specificEpithet, recordedBy, institutionCode, nomenclaturalCode, class.

The dataset contains primary biodiversity data, mostly occurrence data (Figure 7). Some records hold an indirect link to collection specimens. This link is only available in the original database.