Database Reference
In-Depth Information
Current metadata specification allows to create a unique metadata item for the whole set of maps,
assigning the value “series” to the metadata element “ResourceType” (see section 2.2.3, European Com-
mission, 2009). This unique metadata item is used to describe the whole set, since all satellite products
share all their attributes (e.g., geographic bounding-box, lineage, provider, etc.) but time.
In particular, the “temporal extent” element of the metadata describing the series should address the
whole temporal range targeted by the maps: this goal can be achieved specifying a time period either
starting with the first image date and ending with the date of the last image, or declaring a sequence
of time instants, each of them corresponding to an image date (see Table 2). A combination of the two
specifications could be adopted too.
From the metadata provider point of view, managing the map collection like a time series avoids a
long and error-prone metadata creation activity for each map; from the user point of view, the discovery
service presents to her/him a unique resource, instead of a huge amount of similar results. However, in
terms of discovery, it is not possible to filter a single image in the set, on the basis of its own timestamp.
However, the problem of temporal filtering can be solved by filling the resource locator metadata
element “Resource locator” (see section 2.2.4, European Commission, 2009), which could be used to
link the metadata of the series to a Web service providing the maps. In this way, a proper client applica-
tion could allow users to search dates or periods that are automatically combined with the service URL,
and submit a correctly formatted request to the service that provides related maps.
Further Metadata Elements for Temporal Resolution and Events
The spatial characterization of a dataset presents two complementary aspects, i.e., spatial coverage and
spatial resolution; in the same way its temporal characterization cannot be considered complete without
specifying temporal coverage together with temporal resolution, which is also mentioned in the literature
as granularity. An example: a data set contains the daily (time resolution) mean precipitation in a year
(time coverage) as measured by a rain-gauge. Another example is represented by a collection of satel-
lite products such as 16-day (time resolution) biomass maps from the beginning to the end of operation
(time coverage) of a sensor. The temporal resolution of repeated observations is a key element of such
data set attributes as well as it is crucial to assess their appropriateness for specific purposes, e.g., as
related to environmental modelling or monitoring.
Though the current Implementing Rules for INSPIRE metadata (European Commission, 2009) ac-
count for describing both the spatial characterization aspects (see sect. 2.7.2 for Spatial Resolution),
the dual temporal description is missing since there is no element to specify the temporal resolution.
Time resolution can be expressed by the use of ISO 8601:2004 (ISO 8601, 2004), in the form of
Recurring time interval that is specified by adding “R[n]/” to the beginning of a time interval expres-
sion, where “R” is used as the letter itself and [n] is replaced by the number of repetitions, whose unit is
specified in the following part of the same expression (for example R8760/2008-01-01T0:00/P0Y0M-
0DT1H00M, means hourly for the year 2008).
However, it is preferable to provide a separate metadata element, in order to allow for an easier com-
pilation and discovery of metadata. This solution makes it possible to use a linguistic label for expressing
the temporal resolution (such as hourly, daily, monthly, yearly, etc.) instead of a numeric value - the
number of repetitions -, which should be computed on a case-base, and that is not easily interpretable
(as in the example reported). For the use of linguistic temporal specifications in metadata elements see
the following paragraph 3.
Search WWH ::




Custom Search