N^ 3. = UD e A 
- pt J 
\v 
planted with conifers. Some woodlots are small and isolated. 
Squares and rectangulars, patterns normal in ordinary 
forestry, have large cores in proportion to edge but these 
islands’ are too small to develop interior conditions. The 
ecologist locates his new plantations in order to connect these 
isolated woodlots (see upper left corner of Figure 3) in order 
to improve conditions for immigration etc.. He also favours 
plantations were he can get a varied humidity gradient. The 
situation in the left lower corner is the same as in scenario 1 
due to property boundaries. The farmer dislikes remnant 
fields were parts have been covered with forests. The visitor 
has the same comments as above. The new scenario has 
created a somewhat more intimate and varied landscape. 
Mean view-lengths (MVL), based on eight directions, were 
calculated for all open space pixels in the original scene as 
well as for scenarios 1 and 2. These values were classified 
and plots were made. At first sight the original scene looked 
very similar to the two scenarios. The plots are dominated by 
the highest MVL-class. This effect is possibly due to a 
maximum search distance of 1400m. Only pixels that have 
mass pixels quite near on three sides belong to other classes. 
In scenario 1 added woodlots and corridors have cut of parts 
of the upper right corner and hence created a mixed 
landscape with shorter MVL. 
In scenario 2 localisation of new woodlots has not caused any 
great change. The largest effect is to be found in the upper 
left corner were new plantations are screening of an 
'agricultural bay'. 
As a second approach mean view-length was calculated for 
all mass pixels and three plots were made. This gave some 
information about the interdependence of size and form that 
partly could be used in discussions about interior core 
conditions and habitat size. As expected narrow strips and 
small 'islands' showed low values. But if corridors had a 
straight direction the length could compensate and raise the 
value to maximum. 
The area and perimeter of each group of mass pixels and 
open space pixels were calculated. As outcome is dependent 
on grain no transformations to m? and m were made. For 
each of these polygons the A/P-ratio was calculated. The 
total edge length of present forest-polygons was also 
computed. 
In the initial scene 16 forest-polygons were present. Their 
total edge length was 2722, as pixel size was 25 x 25 m this 
equals to 68,050 m. The edge length of each polygon varied 
between 500 and 15750 m. In scenario 1 the total edge length 
of 14 polygons was 3320 (=73,000 m) but in scenario 2 total 
edge length of 12 present forest polygons was lower, 3078 
(276,950 m). This is due to the ecologists wish, he has 
succeeded in creating large areas with interior core 
conditions. 
The A/P-ratio was calculated for every polygon as well as 
maximum and minimum for each area. The quota reveals 
were the patch lies on a scale were maximum is given by 
calculating the A/P ratio for a disc with the same area. 
Minimum is given by A/(A x 2 + 2), which holds for areas 
without islands. Only one polygon lies near its maximum A/P 
ratio. The rest are spread around the mean. 
Scenario 1 and 2 have affected different polygons in different 
ways, but often the A/P-ratio is lowered. Fractal dimensions 
will add more information to the picture. The area dimension 
(1« Da «2) has been calculated for all forested polygons but 
the conclusions are still to be drawn. 
  
Figure 1. The initial scene with two catogories of pixels. 
Mass pixels are shown in black and open space pixels are 
shown in grey. 
97