Vice county

From OpenStreetMap Wiki
Jump to navigation Jump to search

Vice counties are a special category of boundaries which are used for biological recording in the United Kingdom and Ireland. Initially invented by Watson for Great Britain in the 19th century, the concept was extended to Ireland by Praeger in the early twentieth century.

Vice counties were designed to:

  • Have invariant boundaries over time.
  • Split larger counties to provide more comparable units.
  • Merge small counties for the same reason.
  • Separate distinctive biological areas.
  • Retain common boundaries with the civil counties.

There are over 100 vice counties in Great Britain (compared with 86 historic counties), and 40 in Ireland (compared with 32 historic counties).

Boundaries based on detailed digitisation from 1:10000 maps of the Ordnance Survey of Great Britain are available under the OS OpenData licensing terms from the UK National Biodiversity Network. I (SK53) am not aware of equivalent data being available for Ireland. The UK NBN confirmed (in an email to User:EdLoach) that it is OK to use their data in the same way as it is OK to use the OS OpenData (in practice this means that keeping separate boundaries in OSM is not very useful, but see Irish VCs below).

Vice Counties should be mapped using a boundary relation, using boundary=vice_county. In many cases the boundary will follow other features (rivers, coastline, administative boundaries), but when a new way is required the boundary=vice_county can also be applied to that way. They are not administrative units so must not have an admin_level tag. They will only be rendered on specialist maps, and should not be used as proxies for old administrative units, not least because the boundaries are rarely co-incident (e.g., VC50 Denbighshire includes Maelor, whereas Maelor historically is an exclave of Flintshire).

Watsonian Vice counties

Covering England, Scotland, Wales & the Isle of Man. The following boundaries exist.

Users are, in general, advised to use the NBN boundary set (albeit 4 counties have poorly formed polygons), and not to try & replicate the data in OSM. Vice Counties were originally mapped because the NBN dataset had licence limitations.

Vice county Vice county Code Note
Anglesey VC52 relation 360939
Brecknockshire VC42 relation 359909
Caernarvonshire VC49 relation 298872
Cardiganshire VC46 relation 361613
Carmarthenshire VC44 relation 361616
Cheshire VC58 relation 298835
Denbighshire VC50 relation 298843 VC50 includes Maelor
Flintshire VC51 relation 298834 VC51 excludes Maelor
Glamorgan VC41 relation 359902
Herefordshire VC36 relation 380786
Merionethshire VC48 relation 298875
Monmouthshire VC35 relation 359815
Montgomeryshire VC47 relation 298880
Nottinghamshire VC56 relation 181068
Pembrokeshire VC45 relation 361615
Radnorshire VC43 relation 359950
Shropshire VC40 relation 298890
Staffordshire VC39 relation 298889

Irish (Praeger) Vice Counties

The situation is different in Ireland. Firstly VC were not digitised by any official body until very recently, when the [W] National Parks and Wildlife Service commenced a project. Secondly, there are no official open datasets of useful geodata, particularly from the two national mapping agencies: OSI and OSNI. Thirdly, nearly everything which might exist will be limited because it is derived from data/maps owned by OSI/OSNI.

Therefore SK53 embarked on a project to map Irish Vice County boundaries in June 2015. The impetus was attending BSBI meeting in Northern Ireland, and therby meeting 5 of the 40 botanical vice county recorders. By this date, county boundaries on OSM were well mapped for the whole of Ireland: particularly because of the impetus of the Townland Mapping Project. As the Praeger Vice Counties tend to follow administrative boundaries more than the Watsonian ones d; and county boundaries have seen less change in Ireland than in Britain since 1900, the Irish vice counties are rather easier to map accurately than the British ones.

In addition there are a number of papers discussing aspects of delineating the vice county boundaries, notably that of Webb(), which is accessible on line.

Basically Vice County (VC) boundaries were mapped as follows:

  • Where a VC is more or less coincident with a traditional county, the members of the latter were copied to a new relation.
  • Where a traditional county was split, the same process was used, but a crude single way was used to draw the internal boundary.

At this point all VCs were created

  • Places where an administrative boundary had been moved to accomodate the expansion of a town across the river, the VC boundary was restored to the river centre line (e.g., Bray, New Ross, Waterford, Carrick-on-Suir, Limerick)
  • For split counties where railway lines were used for the division (Galway, Tipperary, & Cork) the rough line was split so as to incorporate the railway (by convention the southern-most track was used in multi-track lines).
  • Similarly for split counties where barony boundaries were used, the initial dividing line was quickly refined by using the Memorial Atlas.
  • More complex changes (as outlined by Webb were then addressed providing adequate information was available on OSM). This covered all except the area near the SW corner of Lough Mask & Ballina
  • Existing county borders were checked against the Bartholomew Quarter Inch and the GSGS 4136 1 inch maps, and realigned as appropriate.
  • Some of these were further realigned using GSGS 3906
  • Most detailed alignment involves mapping townlands which are adjacent to County, Barony or Civil Parish boundaries.
  • Ultimately townland boundaries are cross-checked agains aerial imagery & GPS traces to further improve accuracy.

At present (end June 2015), most counties are through the alignment against the 1 inch map (Cork, Tipperary & Limerick are probably the major exceptions). Once this is complete it is expected that the VC boundaries will be accurate to within 100 metres (ie., they will be adequate for the identification of which 6-figure grid squares fall in which VC). Using townlands and features visible on aerial imagery should reduce this discrepancy to something of the order of the accuracy of typical GPSr used by wild life recorders.