Accuracy of the Ordnance Survey Map

Accuracy and the OS Map

The accuracy of the Ordnance Survey (OS) map is influenced by the surveying, data capture and production processes used, by the coordinate projection system employed for the national surveys as a whole, as well as by the detailed specifications for each individual map series. This is useful information to know or have access to, and is provided in support of the guidance provided on the day-to-day tolerances that have been adopted by RoS when working with the OS map.

OS define accuracy in three ways, each of which is described in more detail, below:

• Absolute or Positional Accuracy

• Relative Accuracy 

• Geometric Fidelity

1 Absolute or Positional Accuracy

Absolute accuracy compares the location of a position scaled from a map with the true position on the ground i.e. how closely the coordinates of a point on the map agree with the coordinates of the same point on the ground (in the British National Grid reference system).

The adjoining table represents the absolute accuracy of the OS MasterMap Topology layer applicable to the scale at which the product was surveyed.

Accuracy is often expressed as the Root Mean Square Error (RMSE), which is a measure of the distance from the true position within which about 67% of points would be expected to lie. The maximum expected geometric error on a map is about three times the RMSE.

Certain types of feature, such as road centrelines, vegetation, tidelines, landform limits and underground features, are surveyed to a lesser degree of accuracy. Road centrelines, vegetation and landform limits are subjective and are not surveyed while underground features may have been supplied by third parties.

The Positional Accuracy Improvement (PAI) programme completed by OS in 2006 resulted for a time in two published levels of Absolute accuracy for 1:2500 scale maps. However, the OS no longer make this distinction and instead simply specify one set of values for the whole scale.

Occasionally we receive correspondence from applicants, agents or their surveyors regarding newly surveyed deed plans that have been prepared for an application, and how we reconcile them with the OS map. In many of those instances, the deed plan has been prepared by a surveyor using GPS coordinates (based on the National Grid) so these plans have a high degree of Absolute Accuracy. They often do not, however, take into account the Relative Accuracy of the OS map.

2. Relative Accuracy

Relative accuracy compares the position of a data point or feature on the map in relation to other local data points or map feature.

In general terms, relative accuracy is often more important to map users than the positional accuracy - unless you have a handheld GPS, it is more helpful to be able to see that a wall is 5m to the west of a building than it is to know the national grid coordinates of the wall.

Relative accuracy is normally expressed as a constant value plus/minus an amount proportional to the distance measured. The constant is related to survey practise - the accuracy to which measurements are taken and detail plotted. The adjoining table represents the Relative Accuracy of the OS MasterMap Topology layer applicable to the scale at which the product was surveyed.

The Positional Accuracy Improvement (PAI) programme completed by OS in 2006 resulted for a time in two published levels of Absolute accuracy for 1:2500 scale maps. However, the OS no longer make this distinction and instead simply specify one set of values for the whole scale.

The Relative Accuracy figures are perhaps best understood by using some examples (all are based on the 95% confidence level):

• 1:1250 - When the distance between two well defined points of detail which are 60.0m apart in the real world is compared to the OS map, there would be an expectation that 95% of the time, those points of detail would be represented on the OS map by a scaled distance of between 59.1m and 60.9m (i.e to within +/- 0.9m).

• 1:2500 - When the distance between two well defined points of detail which are 100.0m apart in the real world is compared to the OS map, there would be an expectation that 95% of the time, those points of detail would be represented on the OS map by a scaled distance of between 98.1m and 101.9m (i.e to within +/- 1.9m).

• 1:10,000 - When the distance between two well defined points of detail which are 500.0m apart in the real world is compared to the OS map, there would be an expectation that 95% of the time, those points of detail would be represented on the OS map by a scaled distance of between 492.3m and 507.7m (i.e to within +/- 7.7m).

3. Geometric Fidelity, or the Accuracy of Shape

The principle of geometric fidelity can be described as the ‘trueness’ of features to the shapes and alignments of the objects they represent in the real world.

The OS have a detailed specification for each of their base scales, and these set out for their surveyors and cartographers the degree of generalisation or simplification that is required and is necessary at each scale. It should however be acknowledged that geometric fidelity cannot be as closely defined and measured as Absolute or Relative accuracy, and involves a degree of subjective judgement by the surveyor/cartographer.

Some of the principles of geometric fidelity that are adhered to, dependent on scale, are:

• detail that is square on the ground is represented as square in the data, and shapes must be accurate;

• alignments that are straight in the real world are represented as straight lines within the data;

• lines of sight that pass through ground points should, when plotted at the scale of the original survey, pass through the plan positions of the corresponding points; and

• adjacent features are in sympathy with each other as regards alignment and orientation.