Accuracy: Help Or Hindrance?

It seems to be a truth universally held by neighbours in dispute over their boundaries that a thin line marking a boundary on their deed or Land Registry plan should be reflected by a similarly thin line on the ground. In theory this is correct: a legal boundary is a line or lines of zero width extending from the centre of the earth to some point in the air above. Can we locate this infinitely thin line on the ground to infinitely small accuracy tolerances? The answer is almost definitely no! Particularly when the boundary is based on map evidence, the accuracy of the map means that it can only be marked on the ground, not by a nice thin line, but by a blurred zone within which an infinite number of solutions are possible, all of which have an associated probability of being the “true” boundary. The width of the zone is determined by the accuracy of the map.

Most land is now registered with HM Land Registry (HMLR). The boundaries are plotted onto the relevant Ordnance Survey (OS) map, usually at 1:1250 for urban areas and 1:2500 for rural areas and small towns. These maps fall into several categories:

• 1:1250 scale: maps surveyed at 1:1250 using instrumental methods to provide a framework of controlling detail
• 1:2500 Resurvey/Reformed: maps surveyed at 1:2500 using instrumental methods to provide a framework of controlling detail
• 1:2500 Overhaul: maps originally recompiled from pre-1946 County Series mapping using physical cut and paste to assemble them onto the National Grid
• 1:10,000 scale: maps surveyed at 1:10000 using instrumental methods to provide a framework of controlling detail. Some generalisation of detail.

To complicate matters further, there are two types of accuracy, absolute and relative. Absolute accuracy gives the accuracy of a point to the overall OS National Grid, and is therefore of lesser interest. Relative accuracy is more relevant, since most boundary disputes are concerned with the relationship of a point or feature to the neighbour’s property rather than the rest of the country. It is the accuracy of a point to another point close by. For instance, with the old 1:2500 overhaul mapping absolute accuracy may be poor because of the way in which the old County Series sheets were physically assembled onto the OS National Grid system. However, the 19^th century surveyors were extremely thorough, and for small areas usually achieved good relative accuracy.

The OS has spent £5 million (at today’s prices) over 40 years to establish the accuracy of their work. The absolute accuracy figures they have achieved are given in Table 1.

An example using Table 1 for 1:2500 scale Resurvey/Reformed mapping is that 63% of points can be expected to be within 1.1m of their true National Grid position. Similarly 95% of points will be within 1.9m, and 99% within 2.4m.

However, absolute accuracy is probably only likely to be important if Britain moves onto a fixed boundary system with boundary coordinates determined by satellite positioning systems (GPS), since GPS operates to a global or national reference system. At present, the relative accuracy of features within a small area is more important. The figures are given in Table 2.

Note that relative accuracy is specified up to a certain distance, since this is “local” accuracy. An example for 1:1250 mapping is that, if the distance on the ground between two well-defined points is measured as 60.0m, then there is a 68% chance that the equivalent distance measured on the map would be represented by a value between 59.6m and 60.4m; a 95% chance that it would measure between 59.2m and 60.8m; and a 99% chance between 59.0m and 61.0m. For 1:2500 Resurvey/Reformed mapping a ground measurement of 100.0m would scale at between 99.1m and 100.9m 68% of the time; and between 97.7m and 102.3m 99% of the time.

This returns to the idea of “probability” replacing “certainty” when measurements are taken from map evidence. The client, immersed in his or her boundary dispute, wants absolute certainty in locating the disputed boundary on the ground. The surveyor has the difficult task of explaining that, far from being certain about the boundary position, he or she can only locate it to plus or minus a certain distance at a given probability level!

The situation is complicated even further when one considers that most boundaries, being registered, are subject to the HMLR General Boundaries Rule. This states that “the exact line of the boundary will be left undetermined” (Rule 278, now incorporated into the Land Registration Act 2002). Given the very large number of registered properties, it is not possible for HMLR to precisely determine the boundary in each case. Instead, they carefully edge the property in red up to the nearest line shown on the OS plan (or, if none is available, show the boundary by a dashed line). Since the OS surveys physical features rather than legal boundaries, this line may not be the legal boundary. An example is a hedge and ditch combination: the parcel is edged to the line on the map which is usually the hedge, even though the legal boundary may be the far side of the ditch. Similarly, most plots extend to the centreline of the adjacent road, but HMLR will invariably edge the plot to the feature marking the edge of the road.

It is therefore arguable that the HMLR Filed Plan merely provides a starting point when looking for the boundary, rather than a definitive record of its precise position. Other evidence must be used to determine this on the ground. Principally this will be by close examination of the physical features in the boundary area. Old established fence posts, walls, and hedges are the obvious clues. Where these are well defined the boundary accuracy may be perfect: if the boundary is the face of a building, find the building and the boundary is defined to perfect accuracy. Many features are however less precise: hedge centrelines, particularly old hedges which have spread over the years, can only be defined to a limited accuracy. Old walls are useful, but does the legal boundary run along one face or the other, or down the centre? Generally if existing features can be shown to mark the boundary, the accuracy of boundary location will be fairly high.

The problem comes when the original boundary features have disappeared. Witness evidence may be admissible, but is often lacking in precision. Similarly old photographs may help, but are difficult to measure from. This is where one comes back to map evidence. It is important to find the “root of title”, the conveyance by which the two properties passed into separate ownership. Subject to a variety of legal presumptions, the plan attached to the conveyance taken together with the parcels clause are as definitive of the boundary as it is possible to be. Given the range of quality of conveyance plans from exquisitely surveyed and drawn plans at good scales, to bad hand copies of early OS 1:2500 maps, the problem of quantifying their accuracy is difficult. One way is to digitise the old plans into a computer using a modern high accuracy survey as a reference. This requires well-defined features such as house corners which can be identified on both the old plan and the modern survey. The digitising process will calculate the accuracy of fit of the old plan, and allow the old detail to be superimposed onto the accurate modern survey. By this means a variety of old plans can be presented to the client or the court on a single plan, in a manner which allows the old boundary to be set out on the ground. The method is as objective as possible, is repeatable, and gives a good idea of the accuracy and reliability of the digitised boundary. It is therefore preferable to enlarging the old plans and trying to manually overlay them onto the modern survey on a light table.

As a last resort, if all other sources of evidence have failed, it may be necessary to come back to the HMLR Filed Plan to reconstruct the boundaries by, for instance, digitising. It is after all based on the OS map, which is a record of physical features existing at a certain point in time, plotted to known accuracies.

In summary, the client probably expects that a thin line on the conveyance or HMLR plan should be a thin line on the ground. In theory this is true, but in practice all the boundary expert can do is to indicate a blurred zone of probabilities surrounding a best estimate of the boundary line. He or she can quantify the accuracy of this estimate by using published OS figures, or by using methods such as digitising to calculate the particular accuracy of small areas of old maps by reference to his highly accurate modern survey. However, it may be possible to turn this lack of accuracy to advantage. If the clients can be convinced that the available evidence is simply not accurate enough to resolve the (usually) small differences involved in the dispute, it may be possible to convince them that the only way to solve the dispute is by agreement. It is far better to do this amicably with the neighbour at an early stage, rather than have an agreement imposed by a judge after protracted and expensive litigation.

Nigel Atkinson is Principal of Morton Atkinson Associates, which specialises in boundary disputes, rights of way, and landfill litigation. He is a member of the RICS Boundaries and Party Walls Working Group. He is indebted to the Ordnance Survey for the tables in this paper which have been taken from OS Customer Information leaflet “Accuracy of Large Scale Mapping” 1998 – see their website www.ordsvy.gov.uk.