Geoscience Reference
In-Depth Information
Due to the difficulties associated with using this elegant and satisfactory method,
from an intellectual point of view, we, as scientists tend to use the second method
based on experience, especially if we are working out average values, quantiles
(values that have a certain probability of being reached or exceeded) or even if the
detailed scale at which we want to work exceeds the different scales that are used by
the physical model.
8.3.1.2.2. The empirical approach
In theory, this approach is very different to the previous one, because rather than
giving values to a variable by using determinist laws, the empirical approach is
based on any measurement that is available so that any gaps that exist can be filled
with a value. With this in mind, a large amount of data is required, which is similar
to the information that is being researched (in our example this is wind). This means
that a permanent or temporary network of measurements is also required so that
searches can take place.
At this stage, there is now a choice to operate, between two different approaches:
the first method is based on the use of spatial analysis (spatial interpolation), and the
second method is based on the use of what can be referred to as environmental
analysis. Each of these two methods has been described in more detail earlier in this
topic.
- Spatial analysis leads to the use of spatial autocorrelation between different
points in one area. It relies on the use of Newton's laws of physics, and postulates
the interdependence that exists between these different points is all the more greater
so as the distance that exists between them is small. If we take a typical example, an
unknown value of a particular point which is located half-way between two other
known points (at a speed of 6 and 8 m/second) is equal to the arithmetic average of
the values for these points (in other words 7 m/second). This equation is a non-
linear equation, and this is why most automatic mapping software has a default
setting where the equation is based on the inverse of the distance squared. Carrying
out further studies in this field led to the creation of the field of geostatistics in the
1960s, which was initially used in the world of geology [MAT 63]. The best
applications are the different types of kriging that exist nowadays.
Technically speaking, in order to create the unknown values of the points, it is
necessary to use the values of the points that are known. This is only possible on the
condition that the average area of influence (in other words the range) of each
measuring site is larger than the distance that exists between each site. This is the
main role of the variogram.
If the value of the range is satisfactory, meaning that wind speed can be
recreated, it is recommended to check that no physical or logical factor prevents
such an operation from taking place.
