When preparing my GIS applications session, I began the open-data series with GBIF. I wanted participants to obtain real biodiversity evidence and use it in a spatial project rather than practise only with invented points. The platform’s value lies in bringing together knowledge held by researchers, institutions, museums and recording initiatives and making it more accessible.
What GBIF is and where records originate
GBIF stands for Global Biodiversity Information Facility. It is an international network and data infrastructure, not the organisation that independently made every observation on its portal. Publishers contribute records using standards that allow information to be combined and searched. [1]
My original introduction referred to more than three billion records to convey the scale of available evidence, not to establish a permanently fixed count. For a practitioner, the more important questions concern the records available in the study area, their quality and their relevance to the assessment.
What an occurrence record can contain
Records may include scientific name, observation or collection date, publisher, provenance, coordinates and basis of record. Images and other details may also be available. Some records concern observations, others preserved specimens or institutional collections; they should not all be read as equivalent descriptions of present-day site occupancy.
This connects spatial and attribute data. Coordinates place the evidence; the table explains what was recorded, when and how. Completeness differs between records, so one point’s apparent precision must not be assigned to every other point.
Define a search relevant to the study
The platform supports searches involving countries, areas, taxa and periods through its available options. For a protected area or project, define the search boundary or surrounding scope and explain its purpose. A question about one taxonomic group should not be silently replaced by all organisms returned in the area.
Preserve the query, download date and dataset reference. Search results can change as records are published or corrected. Documentation allows the analysis to be traced to the evidence actually used rather than to an unexplained downloaded file.
Move from download to ArcGIS or QGIS
Available downloads include tabular outputs that can be handled in GIS tools. On import, identify coordinate columns and the appropriate reference, check that points appear in plausible locations and confirm that the attribute table remains linked.
The session exercise used actual occurrence data for a chosen area. Participants could see how an online record becomes a queryable spatial layer that can be related to other evidence. That connects software learning to the origin and meaning of the data.
Review coordinates, dates and duplication
GBIF guidance highlights metadata such as coordinateUncertaintyInMeters and flags issues including invalid coordinates and country-coordinate mismatches. [2][3] Review these with dates, names and basis of record, and consider duplication according to the dataset.
Sensitive species coordinates may be generalised, records may be old, or positional uncertainty may exceed the scale of a siting decision. The sharp point drawn on screen is not automatically a precise contemporary observation at that exact location.
Interpret sparse and dense records carefully
Many points can indicate concentrated recording effort as well as abundant information. Records may cluster near roads and accessible or frequently visited locations, while other areas have been surveyed less.
Missing records therefore do not demonstrate species absence. The data help identify available evidence and knowledge gaps; they do not convert an unrecorded area into one proven to contain no relevant biodiversity. This affects both further survey planning and the conclusions of a desk study.
Connect occurrences with terrain and land cover
Add an appropriate DEM and land-cover product, then check coordinates, scale and reference years. Elevation or surrounding cover can be examined at occurrence locations to investigate relationships with environmental conditions.
This was the core integration exercise: placing wildlife records within terrain and human-activity context. It does not prove that those factors caused the pattern. Record timing, sampling bias and changes between layer dates remain part of interpretation.
What the platform contributes and what follows
The evidence can support desk review, identification of species or locations needing verification, further surveys, research, resource management and protected-area work according to the method. It also demonstrates the value of international open-data collaboration while retaining the need for source attribution and licence compliance.
I encouraged participants to create an account, explore an area and download data as self-directed preparation. The session recording is linked in the references. [4] A useful beginning is understanding what was recorded and what must be checked next, rather than drawing a large conclusion from the first map.
References & original writing
Adapted and expanded from the author’s LinkedIn writing. Technical references are listed below.
