nbul 2004
colour or
image or
e taken to
an extra
yortant for
inimations
interactive
tivity also
the above
oe applied
id be used
], perhaps
d for with
could be
cenario or
und-based
on device
'eing used
2 of other
some sort
to choose
urces at a
ther.
as well as
1 4. If the
|, a greater
tended to
ble in an
phisticated
hould be
s in other
ronmental
pants, and
interesting
ht have a
presented
ss need to
explain its
whelming
n-demand
technical
incertainty
| previous
and shows
:ast of the
ems (GIS)
d land use
enhanced
arieties of
used with
landscape
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
model, in which various ‘cameras’ were placed; Figure | is the
view from one such camera.
Figure 1. The base visualisation used.
The image illustrates the landscape as it may look in 20 years’
time under a particular climate change scenario which includes
efforts to reduce CO2 emissions (wind turbines, solar panels),
changes in crops to reflect differences in climate and water
availability, a reduced water level in the stream, and subtle
changes to the apparent health of non-crop vegetation (trees and
hedges), as well as adaptation and mitigation measures.
Uncertainty in this particular future state comes from numerous
sources, including the nature and magnitude of climate change,
the type and strength of its effects, and the financial, political
and social factors driving policy change and the actions of
farmers. As yet, no specific levels of uncertainty have been
defined, and consequently the modified images are very much
exploratory. As shown here, different elements of the landscape
are given different levels of uncertainty, for example we may be
fairly sure that the stream level will be lowered, but wind
turbines are only a vague possibility due to the planning
process, a lack of feasibility studies for that precise location,
possible local opposition, and so on.
Figure 2. One way in which uncertainty may be shown using
colours
Figures 2 and 3 are examples of colour and clarity modifications
as applied to the image in Figure I. Different colours as used in
Figure 2 could represent different, and relatively easily
distinguishable, levels of certainty, although determining those
levels is a challenge in itself. A legend would need to be
provided with such an image. It can be seen from Figure 3 that
degrees of fuzziness are somewhat difficult to achieve, certainly
more so than degrees of colour, and so this technique is
probably more suited to expressing general uncertainty over a
429
scenario rather than differentiating between specific parts of a
scene. By its nature, blurring is more noticeable on elements
which were previously well-defined, such as the wind turbines
on the skyline, and less obvious on features which already have
a degree of variation edges such as crops.
Figure 3. An example of how blurring may be used to show
uncertainty
In terms of technical feasibility, the images were modified in
PaintShop Pro v7 (Jasc, 2004) using, respectively, an airbrush
tool used on a semi-transparent overlay layer, and a blur filter,
and while they were not time-consuming to produce, it is
evident that it could prove hard to apply such effects repeatedly
with any great degree of consistency. The possibility of
applying such effects automatically currently seems remote,
although Visual Nature Studio does include a PostProcessing
facility which may be of use with certain techniques. Such a
capability would be almost essential if creating animations via
the compilation route, where hundreds or thousands of frames
need to be altered. Post-render colouring of areas of the image
according to uncertainty (which could be specified in the
underlying GIS database) does not seem to be possible at
present, but this may change as the software develops.
¥ View & View 1
* Map (Mecton Bridge. aenal]
© Text information D View?
* Uncertainty (Necton Bridge, around}
& (colour) e View 3
2 (fuzzy! {Grazing Grounds
ei {sound} ground}
Figure 4. A sample layout for an interactive “visualisation
toolbox”
Figure 4 shows a purely artistic mock-up of a possible
“visualisation toolbox” which could be given to viewers to
allow them to toggle different representations on and off,
change viewpoints, or otherwise interact with the display. While
technology already exists to put together user-selectable
displays of still images and other information (such as the web-
browser based system being used in the Kônigslutter area in
Germany (Warren-Kretzschmar & Van Haaren, 2004)), a
working toolbox for interactive visualisations would need to be
