Geoscience Reference
In-Depth Information
Table 14.2
The most suitable method and setting of its parameters to achieve the best quality
DEM for individual relief types (source: the authors)
Ceský Masiv - Czech massif
Karpaty - Carpathians
No. of
input
points
No. of
input
points
Interpolation
method
Interpolation
method
Relief type
Weight
Weight
Plain
Spline with
tension
10
10
Spline with
tension
5
5-10
Flat hilly land
Spline with
tension
20
1
Spline with
tension
10
1-5
Hilly land
Spline with
tension
20
1
Spline with
tension
20
1
Flat highland
Spline with
tension
30
10
Spline with
tension
30
10
Highland
Spline with
tension
20
1-5
Spline with
tension
30
1-5
Flat mountains
Spline with
tension
30
1-5
Spline with
tension
30
5
Mountains
-
-
-
Spline with
tension
15
1-5
Furrow
Spline with
tension
20
1
Spline with
tension
20-30
1-5
Even though the results are contributive, for more objective results it is neces-
sary to continue testing, extend it with other areas, compare the results and observe
whether there is a trend.
14.2.3 Influence of DEM Quality on 3D Relief Expression
Thanks to projection of the earth's surface to the topographic projection plane there
arise deviations both in the relief expression and in its geometrical characteristics.
In case the area of a region or the slope length is used as one of the input quantities
during modelling physical geographic phenomena, we must keep in mind that using
data acquired by 2D measurement is a mistake.
The question is whether the mistake is important for us or not. Table 14.3 shows
that for small values of slope, which can also correspond to low segmentation of ter-
rain, the deviations measured in 2D and 3D are negligible. However, with increasing
slope the value measured in 3D can be up to double; in that case the results of models
of physical geographic phenomena can be influenced more considerably.
The higher the segmentation of the terrain, the bigger the errors. Firstly, there
are the above-mentioned differences in 2D and 3D expression. Secondly, there are
errors in the quality of the DEM which influences the derived parameters.
When comparing deviations of basic statistical characteristics from input altimet-
ric data for example in highlands and flat land, it is clear that deviations (errors) and
differences between the results of individual interpolation methods decrease from
more segmented relief types to less segmented relief types (Tables 14.4 and 14.5).
Search WWH ::
Custom Search