Digital boundaries

Jon Maynard Boundaries Ltd, Boundary Demarcation and Disputes, Rights of Way, Expert Witness, Chartered Land Surveyor


An Internet resource provided free since Febraury 2000 by Jon Maynard Boundaries Ltd

Digital Boundaries in England and Wales

 You are here:    Boundary Problems  |   Appendices  |  Digital Boundaries in England & Wales


Reproduced from Surveying World, Vol 9 No 4, May/June 2001, with the kind permission of the Editor and publishers of Surveying World.

Could the Land Registry's plans to digitise its map records herald a move towards a numerical cadastre? What would be the problems if the UK did decide to follow this route? The author looks at the factors involved and some typical boundary problems.

By Jon Maynard FRICS

Readers of this magazine will know (March/April, page 6 column 3) that HM Land Registry (HMLR) is to digitise its Index Map. Those readers with a fondness for foreign land registration systems may wish to believe that England and Wales has taken its first step on the road to a numerical cadastre. Can they be right? Let's look at the facts.

HMLR operates a register of titles to land and interests in land. The extent of each of 18.5 million registered titles is indicated on its own filed plan, which is based on the largest scale Ordnance Survey (OS) mapping. The Index Map is an index to the filed plans and is currently held on paper as a series of some 400,000 extracts of OS maps which are scattered around the 24 District Land Registries and updated manually.

An important feature of land registration in England and Wales is the general boundaries rule. According to this rule, the extent of a parcel of land is defined b y reference to the physical features that surround it without defining the precise position of the legal boundary of the land. Thus the physical boundary might be a brick wall, but whether the legal boundary follows the centre, or one or other face of the wall, or in fact lies at a small distance from the wall is not recorded.

Why digitise the Index Map?

There are a number of strategic reasons for HMLR to digitise its Index Map. The first is to provide their staff with on-line access to the Index Map to act as an electronic gateway to their definitive title information - the registers and the filed plans - and to allow more efficient processing of registrations and of enquiries. It will eventually be made available for public viewing, and will contribute to the e-conveyancing applications within NLIS.

It's digital, but is it numeric?

Will the vector digitised Index Map meet the needs of a numerical cadastre? Numerical cadastres usually have the following features:

  • fixed boundaries;
  • monuments at every boundary turning point;
  • precise survey of monuments;
  • coordinates of every boundary monument
  • plans showing bearing and distance between boundary monuments.

Fig.1 Example of a Public Index Map

The only one of these features that the proposed digital Index Map will possess is the boundary coordinates, but it would admittedly be easy to derive bearing and distance from these. With neither fixed boundaries nor boundary monuments the digital Index Map is not in itself a numerical cadastre.

What is to stop us from using the digital Index Map as the basis of a numerical cadastre? I would suggest there are four factors:

  • the quality of the information given to HMLR on which first registrations were based;
  • the rules by which HMLR interpret the information they were given;
  • the accuracy of the underpinning Ordnance Survey map
  • the lack of fixed boundaries in England & Wales.

A discussion of these factors follows. The examples given are taken from filed plans, not from the Index Map. It should be pointed out that the Index Map is drawn to a different specification from the filed plans (for example, in cases of horizontal titles there may be only one polygon on the Index map, to which a series of title numbers is applied) and its polygons cannot be taken as accurate definitions of extent of title.

Information given to HMLR

Examples exist of exquisitely drawn conveyance plans at scales as large as 1:360, but often the quality of the conveyance plan leaves much to be desired, as Figure 2 shows.

Sometimes the information supplied to HMLR is not as accurate as it appears to be. The property shown at Figure 3 was sold in 1978 as a semi-detached cottage together with its garden and some surrounding woodland that had not previously belonged to the cottage. The purchaser was required to erect fences through the woodland and the vendor had emplaced pegs to guide the purchaser.

The plan accompanying the vendor's 1978 sales literature, which was a modified copy of the then current Ordnance Survey map, bore the accuracy warning that is reproduced at Figure 4. The plan was re-used as the conveyance plan and retained the accuracy warning. No-one at the time suspected that the vendor's guiding pegs were anywhere other than in the position shown on the conveyance plan, and HMLR faithfully copied the conveyance plan boundary onto their filed plan. The discrepancy was not discovered for 22 years, at which time expert opinion proposed that the land conveyed was the land marked out by the pegs and not the land indicated by the conveyance plan or the subsequent filed plan.

Fig. 2 Plan used, on creation of the parcel shown, in a 1968 conveyance

(Left) Fig.3 Ordnance Survey Superplan Data (in grey) overlaid with the boundary shown on the HMLR filed plan (red) and the author's survey (blue)



Fig 4. Warning note appearing at the foot of the conveyance plan used in 1978 for the registration of the property shown at Fig. 3

Rules of interpretation

The next example demonstrates how the general boundaries rule can deliberately place the red edging on the filed plan in a position that is known to be away from the legal boundary.

The property shown in Figure 5 was created in 1913. The conveyance gives dimensions between four of its five corner points, and the author has confirmed their veracity with measurements on the ground. The filed plan conforms to these dimensions (Figure 5 is based on a photocopy of a filed plan, and this demonstrates how the precise position of the line on the OS map is obscured in the photocopying process, yet professionals routinely attempt to prove points by sending each other photocopies of filed plans).

The 1913 conveyance then goes on to apparently contradict the dimensions it has already given by stating (see Figure 6) that the boundary of the property is actually in the centre of the adjoining road. This statement accords with a commonly used legal presumption concerning the boundary of a property that fronts a road when the road has no identified owner.

Fig. 5 Photocopy of HMLR filed plan for the same property as the conveyance at Fig. 6

Fig. 6 Extract from a conveyance

Thus we can see that the legal and physical boundaries (as shown on the filed plan) of this property are separated by something like 5 metres. In such a case it is clear that the boundary recorded on the filed plan could not be used in a numerical cadastre.

The Ordnance Survey map

In England and Wales the OS map is the base on which all titles to property are registered. It therefore follows that no filed plan can show a boundary with any greater accuracy than is present in the underlying OS map. There are of course a very few fixed boundaries that may contradict the previous statement. There are also a number of boundary agreements usually accompanied by a plan that shows the boundary (as recognised by the land owner on either side) to a higher precision, but these plans are held on the register as separate documents and are not transferred onto the filed plan.

Fig. 7 Ordnance Survey Superplan Data (in grey) overlaid with the boundary shown on the HMLR filed plan (red) and the author's survey (blue)

Many land surveyors claim to find errors in the OS map but these "errors" are usually found to be within the stated tolerance of the map specification for that scale of mapping. Thus, for the 1:2500 scale OS mapping on which Figure 7 is based we should expect to find discrepancies of anything up to 1.2 metres in the relative positions of any two points on the same map. The "error" demonstrated in Figure 7 is about 1.1 metres and is therefore within tolerance. [For an explanation of the accuracy standards of published OS large scale mapping the reader should refer to]

[ The reader should note that this is a dead link and that the information on accuracy standards is now to be found at .]

Taking Figure 7 literally (as one would do if it was part of a numerical cadastral system) the filed plan would have us believe that inside the building the boundary runs within the kitchen of No. 1 rather than along the party wall, whilst in the rear gardens it runs parallel to and to the east of the hedge which separates the two gardens. A visit to the site immediately convinces the observer that the boundary runs in the party wall and along the hedge, regardless of where the map coordinates say it should be. This is because the filed plan (in accordance with the general boundaries rule) uses the line on the OS map, which in turn identifies the party wall and the hedge - even if it shows them slightly out of position relative to other map detail.

Ordnance Survey have announced a rural accuracy improvement programme that will bring the absolute and relative accuracy of their 1:2500 scale mapping to a common standard. Even so, in the best of circumstances - urban 1:1250 scale mapping - we can expect with 68% confidence a relative accuracy between points of +/- 0.4 m. Those property owners who become involved in boundary disputes (still only a tiny proportion of all landowners) are looking for a higher degree of accuracy than this in the positions of their boundaries.

Fig 8. Green circles identify all of the coordinated points used to generate the line dividing the gardens

Relative accuracy is not the only measure of the quality of OS mapping. Consider Figure 8, which shows an apparently straight line, 174 m long, separating the gardens of two properties. All of the points used to generate the line have been marked with green circles.

The cluster of points at the west end of the line is explained by the presence of a garage and two sheds. But what explains the nine points in the centre of the line? At most, the line diverges only 25 cm away from a straight course. Are these intermediate points the result of poor digitising? Or did the digitisers faithfully capture a rather shaky line on the original conventional map? And what would be the implications (within a numerical cadastral system) for anyone attempting to erect a new fence exactly on the line shown on the map?

Fixed or general boundaries

The law in England and Wales does allow for property boundaries to be "fixed", but it is believed that less than 20 registered properties feature fixed boundaries, out of a total of more than 18 million registered so far. In many more instances, neighbouring landowners come to a less formal boundary agreement which is then recorded on the registers of both titles. These boundary agreements are considered to be "general" rather than "fixed" boundaries.

With so few fixed boundaries in the country, some might argue that the physical features used as boundary markers - the walls, fences and hedges - should be used, in lieu of boundary monuments, as fixed boundaries. But:

  • fences usually rot away after 15 years, although the author has come across a house built in 1930 whose original concrete fence posts are still in situ around the garden, but some of them are leaning and not all are in the straight line depicted on the conveyance plans.
  • hedges can spread if untended, and some of them are associated with ditches in circumstances in which it is the hedge that is shown on the map whilst the boundary follows the far side of the ditch.
  • the modern fashion for open plan front gardens leaves many properties with no visible boundaries.

Thus the vast majority of properties in England and Wales do not have boundary markers that could be considered permanent for the purposes of "fixing" the boundaries.

It is worth reminding ourselves of the history of land registration in England and Wales. It began with the Land Registry Act 1862 which required the precise delineation of boundaries. The process of determining the exact position of a legal boundary that was at variance with the physical boundary in use at the time led to many disputes. Only 650 titles were registered under the 1862 Act. Its successor, the Land Transfer Act 1875 removed the need to ascertain boundaries to the nearest inch and gave us our "general boundaries" system.


It should be borne in mind that the examples given above are based upon an examination of the relevant filed plans. These define the extents of individual properties. There are no immediate plans to vector digitise the filed plans.

The Index Map is a different document from the filed plans, drawn to a different specification. It serves to identify the locations of all registered parcels but its polygons are not legally definitive as to extent of ownership.

The digitising of the Index Map should not, and must not, be viewed as a step on the road to a numerical cadastre. It is, and can be, nothing more than a digital image of a record of the "general boundaries" of registered titles in a country that is not accustomed to, and lacks the infrastructure for, a system of fixed boundaries. Furthermore, the cost of providing that infrastructure - precisely determining and monumenting the boundaries of some 25 million properties - and the concomitant opportunities for engendering disputes that it would bring, do not bear thinking about.

Jon Maynard FRICS, a member of the Surveying World editorial board, is in private practice as a chartered land surveyor specialising in expert witness reports for boundary disputes. He is also a member of the RICS Boundaries and Party Walls Working Group.

The opinions expressed in this article are entirely the author's own.

Figures 1, 3, 4, 5, 7, and 8 are reproduced by kind permission of HM Land Registry and of Ordnance Survey and are © Crown Copyright. Media/360/01.


  Return to Top of Page