Scale and Level of Detail

Since mankind exchanged his hunting and gathering existence for a life of land cultivation, enabling him to settle permanently on one spot, he has felt a need to portray the community habitat and surrounding areas in maps. Until recently he was happy with a 2D representation, which worked well, even during eras of ocean voyaging when large areas of the globe were explored, colonised and their resources even exhausted by non-native civilisations. When there was a need to represent the third dimension, such as in hilly and mountainous terrain, lines connecting points at the same elevation above sea level sufficed.

Data Needs
But now the seemingly never-ending growth in world population is intensifying world exploration and drastically changing the mapping scene. Governing such expanding use and coping responsibly and in a sustainable way with associated issues demands large quantities of information. It is a well-known statistic, and one so often proclaimed by so many that it must be true: about 80% of all information used and required by governments contains a geo-spatial component. But this implies not just producing more of the same.

Not More, Different
There is a need for other types of geo-information, a need amplified by the advance of GIS and internet, which allow use of geo-information that differs significantly from the use of paper maps. Coping with current governmental issues requires measures including the integration of various types of geo-information. The officers and decision-makers using this information differ considerably from those of the past, both in number and background. And governing urbanised areas requires highly detailed representations extending into the full three dimensions.

Multi-scale Maps
Resolving all this is thus not just a matter of developing technology to manufacture the same type of geo-information faster and in larger quantities. Technology must also be developed to provide information quickly in the format, scale and at levels of detail requested by the user. It will often be necessary to merge various geo-information sets, for example when a cadastral digital map has to be combined with a road database in designing a road extension. Satisfying such user demands requires multi-scale representations. Scale is a measure of reduction in representations of reality. A map at scale 1:10,000 has a reduction factor of 10,000: an object of 10cm length on the map is in reality 1km.

Nevertheless, in daily use scale is usually associated with level of detail and mapping precision.

Consistency
Various issues are involved in multi-scale representations. Simply reducing the scale by zooming out will result in low-quality map representation: too much detail will be visible. This results in two main disadvantages: (1) the map becomes difficult to read and ugly to look at and (2) transfer of the data by internet is slowed down, even with the broadest band width. Generalisation and selection of objects is therefore required. This means aggregating several contiguous small objects into one; for example individual houses, schools, hospitals and other structures within a city quarter are grouped together in one object typed ‘built-up area’.

Aggregation
Aggregation introduces the issue of semantic consistency; there should be a clear and standardised relationship between objects in the mother database and in derived databases. Objects also have to be represented in less detail. For example, not every curve in a road needs lower-scale representation. This introduces the notion of maintaining topological consistency. An area with several types of buildings located in the mother database on the eastern side of a road should still be on the eastern side after generalisation.

Research
The simplest way to provide maps of various scales is to prepare them beforehand. Scale has to be selected such that the majority of users are served. A main disadvantage here is that all the different representations need storage space. Another solution would be to store just one dataset, the most detailed, from which detailed scales could be derived at user request by making use of appropriate data structures. This would enable the presentation of real-time generalisation on a computer screen without generating new datasets. However, development of the necessary algorithms and data structures poses fundamental problems, some very hard to solve. This sort of map generalisation, whilst greatly in practical demand, is still the subject of research. The tools currently developed by the main GIS vendors approach generalisation via prior storage of databases derived from the mother database.

Synonymous
Before the days of digital storage of map data in a database the map had two functions: as medium of storage and as presentation. The two functions often came into conflict when storing great detail at small scale could make a map difficult to read. The search for a balance turned map-making into both a science and an art. With the creation of geo-information databases the map, onscreen, paper map or otherwise, loses its function as storage medium. Yet when we talk about scale we often mean level of detail. The two terms have become synonymous.