Geoscience Reference
In-Depth Information
Hypsometric layers height steps and colors obey scale-dependent rules of
construction. In small scales steps are usually based on geometric progres-
sion while in large and middle scales steps can be equal as contour inter-
val. Colors should be adapted to the steps chosen, thus in small scales
contrast between neighboring layers will be more prominent and number
of colors will be less. Scale coloring can also change gradually transition-
ing from spectral in small scales to more naturalistic in large ones.
All these suggestions help us to formulate three
basic principles of multi-
scale hypsometric mapping
:
1.
Every map scale should represent relief forms of corresponding size
and hierarchical level with their prominent features and structure lines.
2.
At scales smaller than 1:100 000 intensive generalization should be
made to maintain most characteristic forms of the earth surface.
3.
While changing scales, gradual transition of representation parameters,
such as contour intervals, hypsometric colors and hillshading character
should be provided.
The problems of multiple relief representations and generalization are
closely related to each other. The first studies in automatic relief generali-
zation were based upon filtering (Loon 1978). Since the filtering method
does not take structure lines into account, an adaptive technology combin-
ing filtering and structural generalization was developed by Weibel (1987).
Zaksek and Podobnikar (2005) combine structure lines, characteristic
points and cells of smoothed model as source data for interpolation. Fan
et al. (2007) developed four-stage method using low-pass, smoothing and
threshold filters that are locally combined depending on slope and curva-
ture of the surface. Leonowicz et al. (2009) offer using low and high quartile
filters for valleys and watersheds. The same authors developed curvature-
based algorithm, which is intended for production of small-scale hillshades
(Leonowicz et al. 2010).
Jordan (2007) developed method based upon Strahler ordering of streams.
Basins are derivated for streams of certain hierarchical order and then are
filled by triangulating watershed borders. Ai and Li (2010) use similar
methodology. Process is divided into three stages: stream network gener-
alization using orders, large valleys extension by taking up small ones and
smoothing the surface within the areas of removed valleys. Another inter-
esting field of research is generalization based on Fourier analysis (Clarke
1988) and wavelet transformation (Wu 2000).
