Geoscience Reference
In-Depth Information
appear, particularly in areas such as the Western Pyrenees, where the cubic spline
and kriging methods have produced results that show a lower amount of sunshine
(less than 200 hours) than those produced by the IDW method.
The majority of authors who worked on this research [LAS 94] believe that
whenever a regular sampling grid is used, the cubic spline method tends to produce
results that are equivalent to those produced by the kriging method [DUB 84]. The
cubic spline method is quicker and much easier to use.
The IDW method requires a substantial amount of processing time if there are
several thousand points that been allocated values. The results produced by IDW
interpolations [COL 00] were the most unreliable.
Interpolation methods, which are based on functions where distance is the only
estimator variable, are normally considered as being low-quality interpolation
methods in the world of climatology. It is true, that two neighboring climatological
stations will generally have the same or a similar temperature. However, when a
more detailed analysis is carried out things are not as simple. Temperature, as well
as all of the other climate factors (including sunshine), vary in space and this
depends on factors that are associated with the topography of the Earth's surface or
factors that are associated with land cover.
Such large error estimations can be corrected by the other important aspect of
the interpolation process, which involves carrying out analyses by using continuous
spatial information.
2.3.2.
Interpolations based on continuous spatial information
This second type of interpolation is based on a deductive approach, in which the
factors that produce the climate of a particular area have to be identified. This type
of approach is acknowledged and appreciated by all those people who believe that it
conforms to a certain determinist view of nature, an approach that is used to predict
climate variables. There are many examples based on this approach that have been
published in different papers on the topic.
Certain physical factors obviously need to be considered in this approach and
they include altitude (in accordance with proportionality) which links air pressure
with temperature. Other physical factors include gradient, which controls the flow
of air along the mountain slopes, as well as aspect of the slopes, which exposes the
mountain face to sun rays or to air flow, etc. Other physical factors provide
information to explain the spatial variation of temperature or precipitation: distance
from a particular geographic object, the fractal dimension of features such as
wooded areas, etc. This is why it is common to use a set of factors that have an
influence on the spatial variation of climate factors.
