Bi-temporal foundation for LADM v2: Fusing event and state based modelling of Land administration data 2D and 3D

The prime purpose of Cadastral data – whether in the form of maps, survey plans or notes, or a digital database is the definitive demarcation of the extent of properties – and can be seen primarily as a decision support facility (“Can a structure be built here?”, “Where can I build a fence?”, “Should I buy this property”?). There are, however many additional uses for which this information has been applied – such as a base for the recording of assets such as light poles, underground cables, etc. and as a history of the pattern of land use and subdivision. Although secondary, these uses are important, and should be adequately supported including the historic information. It is a fact that the determination of cadastral boundaries can only be carried out to a certain accuracy, and that that accuracy has been improving over time. Older surveys had been carried out with limited positional control, and using equipment with a low intrinsic accuracy by modern standards, although they correctly represent the topology between properties. As a result, later surveys provide an opportunity to improve the positioning of existing boundaries data without disturbing the topology of the existing data. In addition, engineering works such as road building, can provide a source of high accuracy position data that can be applied to improve low accuracy existing data. This argues that the accuracy of boundaries should be improved in the historic record of the cadastre – after all we would like to see our historic parcels in the position we now know them to have been, so that they are comparable with current boundaries. Likewise, we need to correct inaccuracies in the attributes of the spatial objects and the topology between them (e.g. which spatial units are adjacent to or near a given object). On the other hand, we must not lose sight of the decision-making side of the requirements – so that a past decision can be reviewed in relation to the data as it existed then. If the current knowledge in the database of today is used to review old decisions, they may seem irrational. Data custodians are well aware of this issue, using terms like “update” to indicate a “real-world” change, while using “upgrade” to indicate an improvement of the database representation not accompanied with a change “on the ground”; however database software has not carried this knowledge through – resulting in its loss. This argues for a database with bi-temporal history – where our current best knowledge of the history of the cadastre is recorded, and that history is corrected and maintained, while our past knowledge of the data also recorded as an audit trail (so that we can ask questions like “what did we in 2017 think the definition of this property was in 1994?”). This is realized via two types of time: database (or system) time and real world (or valid) time. The different historic records, combined with changes of datum, can lead to confusion in terminology – where words such as “point”, “position”, “boundary” become overloaded. This paper is intended to provoke discussion of terminology to clear up this confusion, and potentially to assist with an extension of the temporal model as input for the revision of LADM to accommodate bi-temporality.