Environmental Engineering Reference
In-Depth Information
Discussion
The four scenarios were presented to stakeholders, including professionals and
general public, as fixed images as well as fly-over videos. These computer simulations
helped to visualize and characterize the landscape, and are one element, among
others, which can contribute to policy and decision-making. Such techniques have
been used in various similar situations for assessing landscape quality in urban
areas (Cartwright 2008) , at the urban-rural fringe (Lange et al. 2008) or in more
natural landscapes (see, e.g. Appleton and Lovett 2005 ; Bishop et al. 2005, 2008 ;
Ghadirian and Bishop 2008 ; Lange et al. 2008 ; Wissen et al. 2008) . Landscape
visualization has also been used in scenarios with a certain persuasive approach, in
particular concerning climate change (Sheppard 2005, 2006 ; Dockerty et al. 2005 ;
Mansergh et al. 2008) .
The scenarios presented here have been built by a small working group, following
several discussions with stakeholders, and the visual aspect is part of a more global
project. The present work was developed with the objective of being linked to more
comprehensive tools for integrated assessment of agri-environmental policies,
such as the one developed in the SEAMLESS project (Van Ittersum et al. 2008) .
The SLE module has up to now not been integrated in the SEAMLESS-IF, but this
is planned for the near future. Several issues however remain to be solved.
A first issue concerns the landscape that the stakeholder may wish to visualize.
Indeed, the optimum resolution for an adequate visualisation is a cadastral map, but
sufficiently precisely geospatially referenced land-use data are not commonly
available. One way of resolving this problem could be to produce a representative,
neutral landscape instead of a real case study. Some research is under way to produce
such neutral landscapes (Gardner and Urban 2006 ; Le Ber et al. 2006b) , but these
are still insufficiently representative to be used for visualisation purposes.
A second issue concerns the way land-use change can be attributed to specific
parcels or fields. Indeed, models simulating policy scenarios generally consider
statistical data which is most often aggregated at a regional scale (Chapter 7 of this
volume). Disaggregating the results remains a difficult task. Le Ber et al. (2006 a)
suggest the use of a knowledge discovery system based on high-order hidden
Markov models for analyzing spatio-temporal data bases. Taking as input an array
of discrete data, a land-use being attributed to each land unit, a probability matrix
resulting from the scenario assessment modelling chain can then be applied to
produce future land-use.
A large focus has been put on indicators of sustainable development (see for example
Alkan Olsson et al. 2009) . Although the outputs of SLE do not aim to produce any
indicator of landscape quality, the module can be used to visualize some specific
indicators which are relevant at the landscape scale. These indicators can be
presented using more abstract 3D views which do not display the landscape in a
perfectly realistic way, but rather display abstract symbols that show the dispersion of
the indicator over the landscape. In the VisuLands project Wissen et al. (2008)
propose several approaches to integrate indicators in landscape visualisation for
Search WWH ::




Custom Search