Database Reference
In-Depth Information
While the description of the event associated to the different resources can be included in the resource
abstract, there is no way to specify the temporal characterization of the event.
In summary, this chapter proposes to introduce the possibility to include in metadata one or more
events/processes/phenomena of reference for the geodata: a satellite image can be an observation of fires
in a region; a set of meteo records are measures of a rainfall; some thematic maps can be subsequent
representations of urban growth, etc. A temporal extent element should be defined also for the reference
event(s), of course. Notwithstanding further extensions are possible, this chapter proposes to include
the following temporal metadata:
Instant in which the observation/event occurred:
Date, time (e.g. the landslide occurred 11-07-2008 at 8:00:00)
Period of validity, or duration of an observation/event:
period (interval of dates, times) (e.g. the duration of a fire was from 11-07-2008 to 13-07-2008)
Sequence of instances of occurrences/events:
multiple instants of time or dates, periodic or aperiodic (e.g. the dates of the distinct fires
were 11-07-2008, and 12-07-2008)
Sequence of periods of validity or duration of occurrences/events:
multiple periods or durations, periodic or aperiodic (e.g. the times of the distinct fires were
from 11-07-2008 to 13-07-2008 and from 20-08-2008 to 23-08-2008)
Moreover, the following paragraph introduces a framework to allow the definition of imperfect
temporal values in the temporal extent metadata elements regarding both geodata and related event(s).
A FRAMEWORK FOR IMPERFECT TEMPORAL METADATA VALUES
The formats specified in the guidelines based on ISO 19115 and ISO 19119 which regard metadata
(European Commission, 2009) do not allow specifying imperfect values for metadata, both temporal,
spatial, and thematic, which, sometimes, are the only values that metadata producers can provide, given
that they often do not have enough information to define precise values. In real situations, it often hap-
pens that the date or time of occurrence of an event is ill-known. This is due to several reasons: either
because the event was not observed during its occurrence, and thus the date can only be estimated by
experts, or because it occurred long time ago so that its date can be deduced only imprecisely by the
dating methods available (e.g., this is the case of archaeological, geological and paleontological data).
To allow the definition in metadata of imprecise temporal values on the temporal elements of both
geodata and related event(s), fuzzy values should be admitted in the expression of single time points
(e.g., the landslide occurred in the night of 15 th December, 2005), of time intervals (e.g., snow precipi-
tation of three days starting from 14 th December 2008), and of subsequent either periodic or aperiodic
dates/intervals (e.g., fires occurring each summer), i.e., to allow the specification of ill-known instants,
intervals and time series by means of fuzzy instant, fuzzy intervals and fuzzy time series.
First of all a formal framework is described in which all these fuzzy temporal indications are modelled,
adopting the proposal of De Caluwe et al. (1999). In order to express such fuzzy temporal indications
into an XML-type language, TimeML specification language (http://www.timeml.org/site/publications/
specs.html) is adopted.
Search WWH ::




Custom Search