Geoscience Reference
In-Depth Information
- The best approach to be used involves recreating the individual variables that
make up the index. In the example of the Alpes-Maritimes, this would involve
recreating the soil's water reserves, air humidity, and the wind. The water reserve
evolves when it slowly begins to empty (except whenever there is rainfall present).
Thus, the state of the water reserve on a given day D depends on the state of the
water reserve on day D-1. Once the water reserve is full, it needs to be monitored on
a daily basis. On the opposite, the wind and humidity on a given day D are
completely independent of what they were like on day D-1, and this is because these
two variables can change in a few minutes. In order to produce the definitive map,
which shows the index for each point of the grid in an area, successive wind maps,
reserve maps, and humidity maps need to be created. The index is then calculated
for each point by combining all of the different maps that have been created.
- Another approach can also be used to recreate directly the index. This
approach involves using the 23 measurement sites for which the index was initially
calculated. In this approach, it is not necessary to recreate each individual
component of the index.
8.3.3. The results
8.3.3.1. The indirect recreation of the nocturnal meteorological risks, recorded at
03:00 UTC
The most elegant solution involves recreating the spatial fields of the variables
from which the risk index will be calculated.
Figure 8.8 is an example of the recreation of the relative humidity of the air by
using the process of environmental regression and taking the following variables
into account: altitude, relative altitude, slope, aspect, and distance from the sea
(these variables are also used for the other examples of regression which follow in
this section). A detailed examination means that it is possible to check how much
influence the relief of an area and its morphology has on humidity, with a
determination coefficient (R2) of 0.59. The partial correlation coefficients show that
relative altitude has more influence than absolute altitude on nocturnal regressions.
The least humid pixels can be found on the south-facing slopes, especially to the
east of Nice, and areas that are quite far from the sea.
The wind, however, is very difficult to recreate for many different reasons
including the fact that it varies strongly from area to area. And it has weak
relationship and poor statistical correlation with descriptive environmental variables,
such as those that are used to recreate humidity.
Figure 8.9 shows the recreation of the wind field when a different method, other
than environmental regression is used. The method that was used in this case was
spatial autocorrelation, including the process of universal kriging. As is to be
expected, the effects of bumps and hollows are clear to be seen. It is also easy to
Search WWH ::




Custom Search