A program for identification of sites suitable for
wetland projection and restoration has been initiated
for the costal zone of the Laholm Bay (Fleischer et al.,
1991). This program is based on determining the
topographical prerequisites for wetland projection.
Topographical conditions have thus far been mapped
by means of manual interpretation of aerial
photographs or through field work (Wessling, 1991).
These methods are however time-consuming and more
effective approaches are required.
The aim of the work presented in this paper is to
investigate the feasibility of using digital elevation
(DEMS)
information that can be used for identification and
models in establishing topographical
modelling of potential wetlands. The work has been
divided into three parts:
- To develop methods for detection of potential ponds
and their drainage basins through the use of digital
elevation models (DEMs).
- To calculate the areas of the ponds and their
drainage basins.
- To compare DEM-generated pond areas and
drainage basins with manually interpreted areas
(based on aerial photographs).
2. MATERIAL AND METHODS
The study area consists of a 8.0 x 7.5 km area located
north-east of the town of Falkenberg, on the Swedish
west coast. The area is covered by the National Land
Survey of Sweden's (LMV) topographical map sheets
5BNO and 6BSO, scale 1:50 000. Scanned elevation
contours from 1:10 000 topographical maps were
obtained from LMV. The equidistarice of the elevation
contours is 5 meters. All computer work was conducted
using a DEC MicroVAX II. General statistics were
using MINITAB (1985).
applications was performed using UNIRAS (European
Software Contractors, 1985).
calculated Graphical
The scanned elevation contours were first imported
into the ARC/INFO editing facility, where elevation
values were assigned to their respective contours.
These elevation vectors were then used to interpolate
818
elevation values to a gridded digital elevation model.
Since more common and less time consuming
interpolation algorithms (like inverse distance and
Kriging) are not optimal for interpolation of iso-lines, a
spline-based ^ interpolation software ^ package,
ANUDEM (Hutchinson, 1989), was used. The resulting
gridded DEM had a pixel size of 10 x 10 metres and a
resolution in z (elevation) of 0.1 metre. This DEM was
the basis for all subsequent analysis.
A software package developed by Pilesjó (1991) was
further developed in order to identify potential ponds
and their drainage basins and to estimate their areas.
Alltogether, 30 wetlands (ponds and drainage basins)
were modelled in the study area. For the sake of
simplicity, each wetland was assumed to be a pond.
Each potential wetlands' area, volume and drainage
area were determined using this software. The
calculation of areas and volumes was based on a one
meter rise in the water-level of the stream at each
wetlands' location.
2.1 Definitions of Drainage Directions
The first stage of the method of automated drainage
basin detection is to determine each pixel's drainage
direction in the DEM. The eight pixels surrounding
each centre pixel of a three-by-three pixel window
correspond with the eight cardinal points of a compass,
and codes were assigned as follows: N(1), NE(2), E(3),
SE(4), S(5), SW(6), W(7) and NW(8). Each number
represents a drainage direction, and the value 0 is
reserved for pixels with no onward drainage direction.
These can either be sinks, which are defined as pixels
which are lower than all of their eight pixel neighbours,
or pixels located in a flat region. As an option, the
program can eliminate single pixel sinks by assigning
them a new elevation value, equal to the lowest
elevation value of its eight neighbours.
The calculation of the drainage directions is based on
the aspect value of each pixel (Pilesjó, 1991). The
aspects were divided into 45? intervals, resulting in
eight classes, and assigned the codes (0-8), as presented
above. However, in two particular cases it is logically
impossible to assign the drainage direction the rounded
value of the aspect (Pilesjó, 1991):
