Environmental Engineering Reference
In-Depth Information
excursion. Transforming data from one grid to another is a critical job of GIS and a
challenge for the administration of data and their meta-information. A fitting
coordinate system and a fitting projection of the data respectively, is a core issue
of spatial data especially at the boundaries of spatial entities.
b) Spatial Interpolation
Transformation, conversion or the translation of data from one unit or level to
another is the next step of coupling: e.g. transform frequencies to densities, sum-
marize number and abundance of species to the occurrence of communities in
relation to land cover patterns. Most important for the process of integration is
the interpolation of single data-points to area-wide information. Interpolation
creates a surface, a regionalization, from point-data by geostatistical operations.
There are several types of interpolation methods, each with distinct features that
treat the data differently and depend on the special characteristics of the data. When
applying interpolation methods, one must check whether the source data will
change, or the method is subjective (hence, a human interpretation) or objective,
and whether the changes between points are abrupt or gradual. Besides triangulated
irregular networks used to construct digital elevation models, there are other
interpolation techniques such as IDW (Inverse Distance Weighted), kriging (inter-
polate a random field), and spline (approximate complex shapes) which are widely
used (Fortin et al. 2009). In a more ecological context, nearest-neighbour interpo-
lation is a simple method of multivariate interpolation in one or more dimensions.
The final result of an interpolation process is a new layer.
It is common to have interpolated point measurements of annual rainfall (see
Fig.
22.3b
) with IDW. IDW assigns values to unknown points (of an area) by using
values from a widespread set of known points. The value at the unknown point is
estimated by weighting the sum of
N
known values. Most GIS-tools tend to dump
interpolations as a grid-file, bound by given lines like edges of the investigation
area.
c) Spatial Classification
A spatial analysis characterizing distinct areas by information from other layers is
one of the core features of a GIS (see Fig.
22.4
). The values of one layer are
separated by patterns from another layer and then statistically analysed. Thus, the
average character of a larger discrete object can be calculated based on regionalized
data. This function results in new attributes added to the analysis layer.
GIS offer a variety of geo-processing features, all dealing with the management
of the data:
l
Dissolve
aggregates features with the same value of an attribute
l
Clip
cuts features out of a input layer without joining the attributes
