Record Cleaner NBN Atlas Report - Developer Notes

Useful resources and URLS

Mapping NBN Atlas API parameters to DwC terms

Its a bit confusing how the names for fields change, when they change and even the case used. There reasons why will no doubt be historical (there's a reason for everythinbg!). Below is a table of the terms used when generating the report. The resultant CSV uses the DwC terms for the heading. note that all the information in the table below can be obtained from (https://records-ws.nbnatlas.org/index/fields)

Atlas API paramJSON responseDwC termNotes
occurrence_idoccurrenceIDoccurrenceID
taxon_namescientificNamescientificName
taxon_concept_lsidtaxonConceptIDtaxonConceptIDThis is the UKSI TVK
occurrence_dateeventDateeventDate
longitudedecimalLongitudedecimalLongitude
latitudedecimalLatitudedecimalLatitude
grid_refgridReference-Not a DwC term
data_resource_uiddataResourceUiddatasetID
life_stagelifeStagelifeStage

Spatial References: Grid References vs Coordinates

The record cleaner report defaults to using National Grid References (NGR) when available, rather than latitude/longitude coordinates. This is because Record Cleaner expects grid sizes in meters (1, 10, 100, 1000, etc.) for coordinate uncertainty, but the NBN Atlas provides coordinate uncertainties that represent the radius of a circle touching the corners of the grid cell (0.7, 7.1, 70.7, 707.1 meters, etc.) as per the Darwin Core standard. These values don't align with Record Cleaner's expected grid sizes, making grid references the more reliable option when available.

When a grid reference is present, the spatial reference (sref) is formatted as:

{
  "srid": 27700,
  "gridref": "SK1234"
}

The system falls back to latitude/longitude coordinates (SRID 4326) only when no grid reference is available. See the NBN Atlas documentation on grid and coordinate-based records for more details on how the Atlas processes these different spatial reference types.

Back to Blog