Positional Accuracy Improvement - the UK Experience
The emergence of GIS and spatial data capture technologies into government and commercial sectors in the early 1990s meant that asset data were increasingly being associated to the UK's Ordnance Survey (OS) mapping. In the late 1990s, consultations between the OS and its customers confirmed that improved accuracy levels of the existing 1:2500 mapping for rural and rural town areas would be beneficial to all. This was, in the main, due to the fact that the majority of existing 1:2500 scale mapping was based on "overhauled" mapping dating back to the 1950s.
The "overhaul" process involved converting County Series mapping to the British National Grid (BNG) which, in general, resulted in an overall decrease in the Root Mean Square Error (RMSE) absolute accuracy. Over time, further Positional Accuracy Improvement (PAI) changes were introduced at a local level when new map detail was incorporated into the periodical BNG map tile updates. The range of PAI changes varied, with some shifts being small whilst others were significant enough to generate considerable problems; particularly where data depicted legally ratified assets. Overall, the PAI changes, combined with real world change and poor in-house data capture techniques, raised new issues relating to the customer's spatial data quality.
After further focus groups, information papers and seminars (1997-2000) there was consensus that a national PAI programme at 1:2500 scale should proceed. The OS PAI programme was conducted between 2001 and 2006 and delivered improved absolute accuracy for 210 rural towns and associated rural areas, totaling approximately 157 500 km2, across England, Scotland and Wales.
From the outset, the OS and the UK's spatial data custodians acknowledged that the timely introduction and management of the OS PAI programme was crucial. Moreover, custodians believed the transformation of legacy data to meet the OS enhanced PAI standards was more than just a process of shifting asset data to match the new PAI map. It was evident that by maintaining or improving upon the topological relationships and dependencies between all asset data and the new PAI map, the overall quality of asset data could be enhanced.
This paper will focus on the PAI issues faced by Central Government and Local Authorities in the UK. It will endeavour to relate the technical and organisational challenges they faced and outline solutions adopted by major UK data stakeholders.