Geoscience Reference
In-Depth Information
area, as long as there is no discontinuity of more than 200 m between two buildings.
The value of the threshold (200 m) resulting from institutional recommendations,
makes reference to people's space practices. This threshold can obviously be
questioned and takes a different meaning depending on the context. Nowadays,
the approach can integrate new sources. They may be of the object type, such as the
topographic data base of national institutes that include the localization of buildings,
or of the field type (or rather of the “categorical coverage” type to use Plewe's
terminology [PLE 98]) such as the coverages of land use produced by interpretation
of a satellite image. For each new source, the questions of the choice of relationship
measurement (distance, adjacency...) and of the value of the threshold associated
with the discontinuity concept have to be raised. Thus, the city object is the result of
a modeling process, either of composition of lower level objects (buildings), or of
extraction of the “built” category of the land use coverage. Figure 1.3 illustrates this
approach: the construction of such a
fiat
and
composite
object is made from
conceptual reflection (conceptual domain) by operationalizing them through the
implementation of a generic process (model domain) mobilizing observable data
(empirical domain). Depending on the type of selected source, it will be necessary to
assess the capacity of the source to “meet” the model requirements in terms of data.
For example, if the choice to build the city-object is focusing on the extraction of
built patches [GUE 12], it will be necessary to ensure that the resolution of this
source is compatible with the level of resolution associated with discontinuity as it is
defined in the model, namely 200 m. If there is a unique representation of the object
in the concept domain and in the model domain, this representation is no longer
unique in the empirical domain, because it depends on the sources and associated
parameters.
If we continue with the example of the city, this time considering a functional
point of view, namely that of an urban employment center attracting the labour
force of a residential area, then other questions will come up. This design introduces
the concepts of “center” and “area” and the fact that the relationship between these
entities should be measured by the flow of individuals resident in the
area
and
working in the
center.
The model becomes more complex, introducing new objects
and new observables, which multiply the possible ambiguities (definition of a
center, of an area being attracted) and the number of possible choices when passing
to the empirical domain: which indicator to adopt to identify “attracted” entities and
what threshold to associate with it? This is more often specified by coming and
going between the model domain and empirical domain. The separation between
these two domains ensures the possibility to verify the adequacy between the
conceptual object and empirical object. However, it is not without ambiguity. For
example, in the case of the morphological city, the threshold of 200 m, which relates
at the origin to an empirical knowledge, is used in the model domain almost as a
standard, given the generality of the use of this threshold in a variety of countries. In
the case of the functional city, the indicator most often used to measure the intensity
