Project no. 010036
SEAMLESS
System for Environmental and Agricultural Modelling; Linking European Science and Society
INTEGRATED PROJECT Global Change and Ecosystems
D3.7.8.2 Final documentation of model component for Visual Attributes in the Landscape
Due date of deliverable: 31-12-2008 Actual submission date: 18-01-2009
Start date of project: January 1, 2005 Duration: 51 months Organisation name of lead contractor for this deliverable: INRA
Nature of the deliverable: Report
Project co-funded by the European Commission within the Sixth Framework Programme (2002-2006) Dissemination Level PU Public PP Restricted to other programme participants (including the Commission Services) RE Restricted to a group specified by the consortium (including the Commission Services) X CO Confidential, only for members of the consortium (including the Commission Services) SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
SEAMLESS integrated project aims at developing an integrated framework that allows ex- ante assessment of agricultural and environmental policies and technological innovations. The framework will have multi-scale capabilities ranging from field and farm to the EU25 and globe; it will be generic, modular and open and using state-of-the art software. The project is carried out by a consortium of 30 partners, led by Wageningen University (NL).
Email: [email protected] Internet: www.seamless-ip.org
Authors of this report and contact details
Name: Daniel Auclair Partner acronym: INRA Address: INRA, UMR AMAP, (botAnique et bioinforMatique de l’Architecture des Plantes), TA A-51/PS2, Bd. de la Lironde, 34398 Montpellier cedex 5, France. E-mail: [email protected]
Name: Sébastien Griffon Partner acronym: INRA Address: INRA, UMR AMAP, (botAnique et bioinforMatique de l’Architecture des Plantes), TA A-51/PS2, Bd. de la Lironde, 34398 Montpellier cedex 5, France. E-mail: [email protected]
If you want to cite a (P)D that originally was meant for use within the project only, please make sure you are allowed to disseminate or cite this report. If so, please cite as follows: Auclair, D., Griffon, S., 2009. Final documentation of model component for Visual Attributes in the Landscape. D3.7.8.2, SEAMLESS integrated project, EU 6th Framework Programme, contract no. 010036-2, www.SEAMLESS-IP.org , 46 pp.
Page 2 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Disclaimer: “This publication has been funded under the SEAMLESS integrated project, EU 6th Framework Programme for Research, Technological Development and Demonstration, Priority 1.1.6.3. Global Change and Ecosystems (European Commission, DG Research, contract no. 010036-2). Its content does not represent the official position of the European Commission and is entirely under the responsibility of the authors.” "The information in this document is provided as is and no guarantee or warranty is given that the information is fit for any particular purpose. The user thereof uses the information at its sole risk and liability."
Page 3 of 46
SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Table of contents
Objective within the project ...... 7
General Information ...... 7
Executive summary ...... 7
Scientific and societal relevance ...... 8
1 Introduction...... 9
2 State of the Art ...... 11
3 Software methodology and design ...... 15 3.1 Existing software tools ...... 15 3.2 Data needs...... 15 3.3 Data handling and processing...... 18 3.4 Data rendering ...... 20
4 User Manual ...... 23 4.1 Load and parameter setting of the data...... 23 4.2 Export towards SLE...... 24 4.3 S.L.E ...... 25 4.3.1 The SLE Graphic User Interface ...... 25 4.3.2 The land use configuration ...... 26 4.3.3 The 3D visualisation...... 28 4.3.4 The 3D object importer...... 29 4.3.5 The rendering option...... 29 4.3.6 Recording & playing back camera paths ...... 31
5 Illustrative application...... 33 5.1 The situation...... 33 5.2 The scenarios...... 34 5.2.1 Scenario 1: Biodiversity by agriculture ...... 35 5.2.2 Scenario 2: A green city in a Mediterranean forest...... 36 5.2.3 Scenario 3: Urban Pressure...... 37 5.2.4 Scenario 4: The garrigue after energy crisis ...... 38
6 Discussion ...... 39
Conclusion...... 41
References ...... 43
Page 5 of 46
SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Objective within the project
The "SLE" ( "Seamless Landscape Explorer") model component has been developed to provide stakeholders the means of visualising changes in the landscape. This component comes at the end of the modelling chain of Seamless. It has been designed to import scenario- based landscapes that are generated with the help of the farm and environmental process models in Seamless (i.e. FSSIM). The results can then be used either individually by the policy-maker as a decision tool, or as a support for discussion and negotiation. The users can explore the landscape, find environmental issues that are relevant, and discuss their opinions to form strategies for resolving them.
General Information
Task(s) and Activity code(s): T3.7: A3.7.5 Input from (Task and Activity codes): T3.3, T4.7 Output to (Task and Activity codes): N/A Related milestones: M3.7.5 Software download and user manual available at: http://umramap.cirad.fr/amap2/logiciels_amap/index.php?page=sle
Executive summary
This Deliverable provides the documentation for the module which has been developed to assess the visual attributes of landscapes. One important aspect of cultural landscapes concerns the perception by the public of landscape elements and changes in the landscape. This is addressed here by developing a model component specifically targeted at visualising changes in the landscape. This component comes at the end of the modelling chain of Seamless. The input data for such visualisation concerns terrain data (DEM), land cover (present land cover, and land cover resulting from the outputs of FSSIM), and a library of detailed textures and vegetation models. The outputs are qualitative and are to be used in the post-treatment analysis, and/or in the negotiation phases. The present document accompanies the "SLE" software for visualising landscapes in 3D, which has been developed following the procedure and methodology described in the project deliverable PD3.7.2 and in deliverable D3.7.5. The present module has been developed as a stand-alone desktop application, and an illustrative application is presented.
Page 7 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Scientific and societal relevance
Three-dimensional visualisation of the landscape is often used for communicating with various stakeholders. Static, web-based landscape visualisation tools have made considerable progress in recent years, such as for example Google Earth, covering the entire planet in 3D. Such visualisations are based on aerial (satellite) imagery, at a specific date, but are not dynamic. The challenge in the Seamless project is to view future changes in land use, according to scenarios. The pressures causing changes in landscape may come from the FSSIM model, or from any other scenario not necessarily modelled through Seamless, which will then be translated into changes in the spatial configuration of the landscape. The mapped results (environmental data such as land cover and land use) are specific to some particular regions and should be available through a DataBase sufficiently detailed for the region of interest. They are used here to compute and visualise a 3D scene. Although land managers and policy-makers generally have a good experience of what result can be expected from their decisions, they are often faced with difficulty when trying to communicate the visual impact of a future management option to all the stakeholders (local and regional decision-makers, land managers, landscape planners, and various communities involved in outdoor activities). With the present module, the landscape visualisation component is launched at the end of a scenario simulation to allow for exploration of landscape changes. Visualisation could have a significant implication for the choice of effective land-use policy, and could be used as a basis for discussion and negotiation within the community.
Page 8 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
1 Introduction
The present document accompanies the "SLE" ("Seamless Landscape Explorer") model component and software designed for visualising landscapes in 3D, which has been developed following the procedure and methodology described in the project deliverable PD3.7.2. and in deliverable D3.7.5. This module has been developed as a stand-alone model. It can therefore run independently from the other SEAMLESS components. It has been developed in this way, following the first discussions and workplans. The inputs can come from the Seamless modelling chain, or from any other project interested in landscape change. In the following parts of this document we first provide (section 2) a brief state of the art of landscape visualisation models. We then provide in section 3 the software methodology and design. Section 4 gives the outline of a user manual. Section 5 presents an illustrative application, based on four scenarios applied to the Pic Saint-Loup area in Mediterranean France. The document ends with a discussion in section 6.
Page 9 of 46
SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
2 State of the Art
Rural land managers, foresters and farmers, but also local decision makers, local authorities and members of local governments, are increasingly aware of the necessity to take into account the perception of the landscape by the general public, and to predict the evolution of landscapes according to management decisions (Bergen et al., 1995, Bell 2001). Different management choices can lead to similar, or to very different landscapes. The positioning of woodlots, of fields, and of agroforestry areas, the type of silvicultural management (selective or systematic thinning, artificial pruning, clear-cut or shelterwood systems, reforestation, choice of species, etc…) or agricultural system (rotation, land attribution, crop allocation, etc…) and the balance between forest and agriculture, are susceptible of drastically modifying the visual aspect of the landscape. The public generally considers the landscape as timeless, and resistance to any change is often very great. Forest management can arouse public antipathy and in some cases can lead to outspoken criticism. Conversely, a well thought out management plan can considerably improve the visual aspect of a rural landscape (Savill et al ., 1977). Although land managers generally have a good experience of what result can be expected from their decisions, they are often faced with difficulty when trying to communicate the visual impact of a future management option to all the stakeholders (local and regional decision-makers, land managers, landscape planners, and various communities involved in outdoor activities). The perception and aesthetic evaluation of a landscape can be analysed according to three organisation levels: an objective dimension (formal landscape organisation, proportion, composition rhythm of various biophysical elements), a cultural dimension (related to social groups, to history, to representations of nature), and an individual subjective dimension (Sauget and Depuy, 1996). Some socio-economists have undertaken to assign values to the aesthetics of a landscape (Thomas and Price, 1999), but this subjective dimension generally leads land managers to seek a consensus between stakeholders, by negotiating around some kind of visual representation (Tyrväinen and Tahvanainen, 2000). Representing a landscape has always been a difficult task. As early as the 16 th century, “bird's-eye" views sought to combine two-dimensional maps with a representation of perspective. Nowadays, land managers have a number of tools at their disposal (Perrin et al. 2001). • Maps and plans have often been —and still are— used for their rigour and for the possibility of quantification they offer. • Their use is now greatly facilitated by geographic information systems (GIS). These computer tools include georeferenced databases, which help to produce two-dimensional maps. They contain large amounts of information, which can help to represent land-use and its evolution. GIS manufacturers provide three-dimensional visualisation techniques, which however remain too restrictive to represent satisfactorily large landscapes: the terrain is generally well represented, but particular landscape features and architectural elements are often simply extruded, and do not represent real volumes. • Photographs can be used directly, or can be modified with the help of computer-aided imagery. This technique has been used for example by Tress and Tress (2003) to discuss various scenarios with stakeholders. It is however very time-consuming to build different scenarios of future land-use. • Schematic representations can be produced simply by drawing sketches or diagrams, either by hand or with the help of computer systems.
Page 11 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
• Photographs and drawings have no direct link with maps, and only show a limited number of viewpoints. They are purely visual techniques, which cannot easily simulate the reactions of a landscape to human interventions. • Virtual imagery is a modern tool which can help represent landscapes, in three dimensions, and which can also provide the possibility to include a dynamic aspect. Three-dimensional visualisation of the landscape provides means that are better understood than maps, especially for the general public. With such methods, visual changes of the landscape can be shown very impressively, which can allow for an intuitive assessment of the visual landscape quality. Static, web-based landscape visualisation tools have made considerable progress in recent years, such as for example Google Earth ( http://earth.google.com/ ), covering the entire planet in 3D. The French geographic service proposes Géoportail ( http://www.geoportail.fr/ ), a tool which couples 3D visualisation with a number of additional GIS layers, such as high quality maps, as well as extruded buildings. A number of companies offer facilities for creating and customizing maps and "virtual visits", such as the Microsoft product "Virtual Earth 3D", a plug-in for 3D visualisation ( http://www.microsoft.com/virtualearth/default.aspx ) or Skyline Globe http://www.skylineglobe.com . Google Earth offers facilities with Google Earth Outreach ( http://earth.google.com/intl/en_uk/outreach/index.html ) and SketchUp (Google Inc 2006). A "Google Earth Community" ( http://bbs.keyhole.com/ubb/categories.php/Cat/0 ) and several international societies have appeared (for example the International Society For Digital Earth http://www.digitalearth-isde.org/ ). Examples are shown below: Figure 1 shows examples of 2D visualisation with the French Geoportail and with Google Earth for two different situations, and Figure 2 shows details for Flevoland in 2D and in 3D. Figure 3 shows an example of 3D view of a coastal village, with 3D extruded buildings (Géoportail). Figure 1a: Example of a landscape viewed from above: the Restinclières agroforestry estate (France) viewed with the French Geoportail
Page 12 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Figure 1b: Example of a landscape viewed from above: the Flevoland area (Netherlands) viewed with Google Earth
Figure 2: Details of the Flevoland area (Netherlands) viewed from above in 2D (left) and in 3D (right), using Google Earth. Note the difference in colour for some of the details, which is due to different dates for the aerial photographs
Figure 3: Example of a 3D view of a coastal village (Collioure, France), with 3D extruded buildings (Géoportail)
Page 13 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
The above examples show static views based on aerial (satellite) photographs, at a specific date, but are not dynamic. The challenge in the present project is to view future changes in land use, according to scenarios. It is also expected to produce more details than can be seen in the above Figures. In the last 20 years, research on 3D landscape modelling has increased considerably, mainly for urban planning, but also for studying rural and forest landscapes (Danahy 1989, Auclair et al., 2001a, Bishop et al., 2001, 2003, Lovett et al., 2001, Orland et al., 2001, Herwig and Paar 2002, Snyder 2003, Dockerty et al., 2005, Paar and Clasen 2007, Lange et al., 2008). Several international meetings have addressed the issue of visualising landscape, for example "Our Visual Landscape" (Ascona, Switzerland, 1999) or "Futurescapes" (Belfast, UK, 2002), resulting in special issues of the journal Landscape and Urban Planning (Lange and Bishop 2001, Lovett 2005). Specific software have been developed (Perrin et al., 2001, Bishop et al. , 2005), which increasingly benefit from recent advances in technology, such as "visioning hubs" or "immersive tools" (Sheppard 2006, Salter et al., 2008) or "augmented reality" (Ghadirian and Bishop 2008). Some freeware and/or open-source software have been developed in addition to those proposed by commercial companies (for example Geomantics in the UK which proposes freeware versions or more complete professional versions, http://www.geomantics.com/ or LandSIM3D ® commercialised by Bionatics in France http://www.bionatics.com/ ). Many are linked to GIS, and have been described in several literature reviews (Ervin and Hasbrouck 2001, Appleton et al., 2002, Pettit et al., 2008). Often specific 3D objects can be included in the scene, such as vegetation (Auclair et al,. 2001b, Muhar 2001), and several approaches have been described in order to integrate indicators into landscape visualisations (Bishop et al., 2008, Wissen et al., 2008). A particular emphasis has been put on reliability and validation of landscape simulation (Sheppard 1982, Daniel 1992, Lange 2001, Bishop et al., 2005). Several European research projects are involved in the issue of landscape visualisation, such as VisuLands (Wissen et al., 2008), Greenspace (Lange et al., 2008), Sensor (Helming et al., 2008), BioScene (Soliva and Hunziker 2008). These projects however address different issues (rural vs urban landscapes), at different scales (country, region, "territory"), and although some aspects are common, it appeared necessary to develop separate components, although on a common basis. In the present document, we describe a landscape visualisation freeware component, which is part of the SEAMLESS project. This component can be launched at the end of a modelling chain destined to simulate a particular agricultural and/or environmental policy, to allow for exploration of landscape changes. Visualisation could have a significant implication for the choice of effective land-use policy, and could be used as a basis for discussion and negotiation within the community. The pressures causing changes in landscape can be simulated by a bio-economic farm model, such as FSSIM. This can then be translated into changes in the spatial configuration of the landscape. The mapped results (environmental data such as land cover and land use) will be specific to each individual region, and should be available from a GIS Database. They will be used here to compute and visualise a 3D scene. In this deliverable we first describe briefly the software methodology and design and the data processing, and in a second phase provide an example of application based on a study of four scenarios in the Pic Saint Loup area of the French Mediterranean region.
Page 14 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
3 Software methodology and design
To build a system that computes a virtual landscape from accurate spatial data and delivers a reasonable realistic representation of an existing landscape, a large amount of data is necessary. Firstly, we need to build a real-time renderer that is powerful enough to handle large datasets of landscape terrain. Secondly, a matching virtual representation of an existing area has to be constructed, including vegetation and man-made structures. We have built a system using a free graphic engine as a base for building a virtual environment. To obtain a suitable 3D dataset of an area of interest, users of the system can select an area on a 2D digital map in a GIS and the application sends the matching data to the landscape visualisation component. The generation of the 3D model and its rendering is done by this software component, that we have called here "Seamless Landscape Explorer" (SLE). Thanks to this virtual environment, a post-model analysis can be conducted to address issues specific of the landscape, with an objective of participative planning and negotiation with the various stakeholders. The challenge of the SLE software is, without proprietary tools or database preparation, to extract specific data (elevation, imagery and land cover) from a GIS, fuse this data in a procedural manner to enhance its apparent quality, and add vegetation objects to the scene.
3.1 Existing software tools
Today, 3D landscape visualization software such as Vue ® (E-on Software, http://www.e- onsoftware.com/ ), or World Construction Set ® (3D Nature, http://3dnature.com/ ) are available. They provide the possibility for modelling and rendering landscapes with a high degree of realism, but are specifically designed for infographists. Since 2000, Visual Nature Studio ® (3D Nature) is a GIS-compatible version of World Construction Set ® and in 2003; an add-on has been released that introduced real-time capability on a lower level of detail. Furthermore, our partner Bionatics ( http://www.bionatics.com/ ), specialist in 3D plant modelling and landscape design, has released LandSIM3D ®; a simulation and design software for quickly modelling existing landscape in 3D from GIS data sources. It permits to import a project, visualize its integration in a real site, study its alternatives, its future evolution, and decide for the best options. These software tools are actually used for landscape planning. However, these are commercial software and not designed especially for research scientists. That is why we developed a free GIS-compatible and interactive landscape visualisation software, based on available open-source components.
3.2 Data needs
The representation of 3D landscape models requires a variety of components and corresponding spatial data types. These include terrain texture (orthoimagery, raster maps), digital height models (DTM, Digital Terrain Model or DEM, Digital Elevation Model, a digital representation of ground surface topography or terrain; or DSM, Digital Surface Model, a topographic model of the reflected surface of the earth that can be manipulated by computer programs), vector-based 2D geo-objects, 3D objects, object textures, animations and hyperlinks. The spectrum of these components ranges from very large spatial objects to
Page 15 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009 large numbers of complex and possibly dynamic 3D objects. These data types have very different characteristics and requirements in terms of management, visualisation and multi- scale representation.
• Digital terrain elevation data (raster) At present, available through many databases, the GTOPO30 elevation data (around 1:1 000 000 scale ( http://edc.usgs.gov/products/elevation/gtopo30/gtopo30.html ) are not accurate enough. For testing, we have used the NASA Shuttle Radar Topographic Mission (SRTM) data. This data is currently distributed free of charge by USGS and is available for download from the National Map Seamless Data Distribution System, or the USGS ftp site (http://glcfapp.umiacs.umd.edu:8080/esdi/index.jsp ). The SRTM data is available as three arc second (approx. 90 m resolution) DEM. A one arc second data product was also produced, but is not available for all countries. The vertical error of the DEM is reported to be less than 16 m. We need at least these kinds of scale datasets but more accurate data like the IGN ® BD ALTI ® (50 m resolution, https://professionnels.ign.fr/ficheProduitCMS.do?idDoc=5323461 ) gives better results.
• Land covers classification data (raster or vector) For vegetation placement; we can use self-made scientist datasets and/or CORINE Land Cover like datasets ( http://dataservice.eea.europa.eu/dataservice/metadetails.asp?id=822 ). Land use and land cover maps show areas of land as 'parcels' or polygons. Each parcel has attached to it a list of values or attributes, covering such topics as land cover class, parcel area, length of boundary, processing history, knowledge-based correction and identification of the original satellite scene. We need these kinds of data to display actual and simulated landscape, i.e. the environmental impacts of a simulated scenario in Seamless need to be translated into land cover data. We will further discuss about the data translation in section 6.
• Geotypical textures library (raster) For “texture splatting” (a method for combining different textures) on the terrain, we need a collection of textures for the different land cover classes we could find (e.g. wheat, corn, grassland…). A preliminary library of textures is being developed, and if necessary, according to the user requirements, a variety of different textures can be added.
• Orthophoto (raster) For rendering more realistic landscape, an orthophoto (an aerial photograph that has been geometrically corrected such that the scale of the photograph is uniform) can be draped over the terrain and blended with the geotypical soil textures.
• Vegetation model library Vegetation elements, especially trees, can render the landscape more realistic than simple 2D textures. Producing high quality 3D vegetation models is a very time-consuming and costly operation, and it is not yet planned to create such a library for vegetation objects. As a viewer moves away from the object, we need the model to switch from a geometric representation to a cross-polygon model, and to a billboard (one or a combination of several two-dimensional images), such as presented in Figure 4. Building a variety of realistic, multiple level of detail plant models could be accomplished using a package from our commercial partner ( i.e. REALnat ® by Bionatics ® http://www.bionatics.com/ ). An example is presented in Figure 5.
Page 16 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Figure 4: Two level of detail plant models from REALNat ® by Bionatics ®
A high level of detail is necessary when visualizing the object from a close position (left) and a lower level of detail when visualizing it from afar (right) Another solution is to represent less realistic plants using basic primitives such as cone, cylinder, sphere or convex hull (Figure 5). Figure 5: Examples of a very detailed plant from the Bionatics ® library(left) and simple plant forms (right) built under the GEOM module of AMAPmod (Boudon et al., 2001), now integrated as PlantGL in ALEA (Pradal et al., 2004) http://gforge.inria.fr/projects/openalea/
At present, SLE is available only with a small number of simplified samples of AMAPsim files (full detailed 3D tree models) and free 3D models from 3D-Diggers ( http://www.3d- diggers.de ). Furthermore, we have planned to build a content library based on existing vegetation textures and models. This library should eventually include some major European species and be linked to the land cover classes. Such a library should also offer the possibility for the user to include additional textures according to land-use and vegetation not accounted for in the first versions.
Page 17 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
3.3 Data handling and processing
The system described here is able to automatically generate suitable 3D content models from spatial 2D data. SLE users have the option to load geospatially referenced data resulting from different policies and select an area of interest for the visualisation. Ideally, this should eventually be possible anywhere in Europe, if the appropriate data (see previous section) is made available. To manipulate these data we propose the user friendly Open Source Geographic Information System (GIS) Quantum GIS (QGIS, http://www.qgis.org/ ). A 3D data conversion tool has been developed and integrated into QGIS. It works as an external module of the visualisation one and converts raster layers and shape files into a format the renderer can read. The 3D visualisation module has been developed as a stand-alone model. It can therefore run independently from other modelling components and be launched by the GIS. Currently, the link with bioeconomic models is done by database server connection. In Figure 6, we describe the flow of data through this spatial database and SLE. Figure 6: The Seamless Landscape Explorer diagram
POSTGRES Database SERVER Spatial database
Quantum GIS Additional GIS data : e.g DTM, Landuse field shapes, Orthophoto SLE Exporter Plug-In
Project xml file + Formated data
SLE Project configuration file Textures and Camera bookmarks file 3D objects Video path bookmarks file Editor Renderer (plants, trees, houses, etc)
Real-time rendering Screenshots Videos
Page 18 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Land-use distribution can be downloaded from the POSTGRES spatial database by QGIS. More accurate data such as field patterns and a digital elevation model must be also loaded (from the local user hard drive or from a server) to allow SLE visualisation. The SLE QGIS plug-in can allocate each agricultural parcel to a specific land-use. This will be done by importing the proportion of each land-use class computed by the bio-economical farm model and distributing it on the field pattern according to specific allocation rules. At present the rule implemented in the model is simply random allocation of changes, but this should be improved in future versions. Then the plug-in can export any extent of approximately 10 km by 10 km selected by users. Data are cropped, fused and formatted to be visualized. An xml project file with geo-reference and land-use information is also written. SLE is then launched and the user can edit the geo-typical configuration (texture and vegetation) for each land-use class. In practice, the land-use geo-typical parameters consist in the following items: • A tileable texture representing typical ground cover. This is used to build the texture that is draped over the terrain when rendering. A number of examples are shown in Figure 7. • A list of possible objects in the land use type defined by the following parameters : o The type of 3D objects: it is possible also to define objects like rocks, as long as they have a natural distribution. o The type of spatialization: the object’s distribution over the terrain can be done randomly or by rows. o The level of detail: the object can be displayed in 2D (billboards) or in 3D according the distance to the camera. o Object density: this is used in conjunction with a random offset to determine the number of objects in a specific area. o Inter-row/Intra-row spacing: this is used in conjunction with row spatialization to specify the distance between objects. o Scale variation information: this is used to generate variations in the appearance of objects, and is particularly useful when billboards are used to render trees. Once the configuration has been defined, the terrain mesh, the textures and the vegetation objects can be computed and the user can explore the landscape in real-time. It is possible to bookmark some interesting points of view or camera paths, take screenshots, modify the terrain, add 3D objects manually and modify the rendering. Several "test case landscapes" are now under study. The option which is at present being investigated is to select only a small number of "representative" landscapes for which sufficient spatially explicit data are available, and concentrate on these. One important issue concerns the way to disaggregate and spatialize data which have been aggregated for running farm models.
Page 19 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
3.4 Data rendering
Technically, a dynamically optimized elevation mesh from the digital terrain elevation raster is first computed using the Geographic MipMaps technique (MipMaps are pre-calculated, optimized collections of bitmap images that accompany a main texture, intended to increase rendering speed and reduce artifacts, de Boer 2000). Then the mesh is textured with the “texture splatting” technique (Bloom 2000; Tyrväinen and Tahvanainen 2000) and with satellite imageries or thematic maps. Texture splatting means that from a set of appearance parameters, a selection of tilable textures is blended together and then splat onto the surface. On Figure 7, we can see four possible types of ground cover for landscapes (top). These are large scale layer ground covers and were obtained from aerial photographs. The middle four textures define ground cover on small scale showing grass and soil. In the bottom, we compare an orthophoto of the Pic Saint Loup (South of France) with texture splatting according to the land-use map that forms a pseudo orthophoto. Figure 7: Examples of texture splatting
The number of 3D objects on the landscape is very important. For example, if the landscape represented is one kilometre squared and with an average density of one object every ten metres squared we would have 100,000 objects on the landscape. This is more than many systems can handle in real-time. It is therefore necessary to give some form of organisation to the 3D scene.
Page 20 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Thus, we established a fixed grid around the camera to manage the vegetation data for each layer of plants and other natural objects. Each grid cell contains all of the data to render its layer in the physical space it occupies. For each layer, we establish a distance from the camera that the layer needs to generate visuals; this determines the size of our virtual grid. This operation is done in real time and care must be taken to ensure that “planting” is a fast operation. The different layers of vegetation consist in trees, shrubs, small plants, rocks, and other small objects to complete the illusion of natural complexity. We apply random transforms to vary their size and orientation as we pick our “planting” points. Some of these can be represented as 2D textures on 3D planes (billboards) just as grass is, but the richness of the environment is enhanced when we mix an assortment of geometric objects, as well. An extension to the “billboard” technique has been developed, the impostor rendering (Day and Willmott, 2005). Impostors are simply dynamic billboards; this means that the texture is updated dynamically according the viewing angles, so as to reduce the visual error incurred by using a flat object (Figure 8). Finally, the user can activate options of rendering to add realism and immersion to the landscape such as: sky, clouds, sun, haze, shadows, lake and stereo-vision. Additional technical details are described in Griffon and Auclair (2009) Figure 8: Impostors of a 3D AMAP plant (Pinus halepensis)The same 3D plant is visualized here from 16 different directions
Page 21 of 46
SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
4 User Manual
Seamless Landscape Explorer (S.L.E.) is a software that computes a virtual landscape from accurate spatial data and delivers a reasonable realistic representation of an existing landscape. To realize this visualization, it is necessary to follow two successive stages allowing to load and parameterize and to export the data which will be used by the renderer. To download SLE : http://umramap.cirad.fr/amap2/logiciels_amap/index.php?page=sle
4.1 Load and parameter setting of the data
To manipulate the data we propose a user friendly Open Source Geographic Information System (GIS) named Quantum GIS (QGIS). QGIS allows you to create, visualise, query and analyse a variety of spatial data types including terrain texture (orthoimagery, raster maps), digital height models (DTM, DSM), or vector-based 2D geo-objects. The graphic interface of QGIS appears in the following way: Add Vector layer Add Raster layer Add PostGIS layer
Data layers
Load a QGIS project Save a QGIS project Toolbar Map View All this information can be found in the User and Installation Guide of QGIS at: http://download.osgeo.org/qgis/doc/manual/qgis-0.9.1_user_guide_en.pdf
Page 23 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
To use SLE correctly, you first have to load in QGIS at least one landuse layer (with landuse attribute table) and one digital elevation model (DEM). You can also optionally add orthophotos. Then, you have to rename every layer so that it is correctly exported towards SLE (right click on the layer: Rename). • Rename the land use layer : landuse • Rename the DEM layer : terrain • Rename the orthophotos layer : ortho You can now save your project in QGIS.
4.2 Export towards SLE
The SLE software allows to choose exactly the zone which we wish to see represented. For that purpose, it is necessary that the zone of display of QGIS (Map View) corresponds to the area which we wish to export. When this choice is done, you have to open the menu "Plugins" and select “SLE Exporter” then “Export 3D landscape”.
A popup window appears and allows you to choose which specific attributes of the land use layer you wish to visualize. Landuse combo box
Page 24 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Select a land use attribute data from the ComboBox, enter a name for the SLE project and click on compute. Data are automatically cropped, fused and formatted to be visualized. An .xml project file with geo-reference and land-use information is also written. Next, SLE is launched and the user can edit the geo-typical configuration (texture and vegetation) for each land-use class.
4.3 S.L.E
4.3.1 The SLE Graphic User Interface
1
2
3
The graphic interface of SLE can be divided into three different parts: the menu bar (1), the 3D view (2) and the SLE configuration (3). The menu bar (part 1): • File: - Load project (Ctrl-P): load a SLE project. - Quit: quit SLE • Edit: - Full screen (Ctrl+F10) : show the 3D view in full screen • Help: - About : the credits Page 25 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
The 3D view (part 2): You can navigate in 3D with the keyboard: • Left control = Move left • Right control = Move right • Up control = Move forward • Down control = Move backward • PgUp = Move up • PgDn = Move down • Right Click + Cursor Right = Rotate right • Right Click + Cursor Left = Rotate left • D = Render terrain solid/wireframe /point • F = Anisotropic filtering On/Off • F1 = Takes a screenshot (placed in the current directory) • Insert = Record the camera path. • Return = Play the recorded camera path. • Space = Record keyframe camera position. • [ or ] = Change frequency between two keyframes.
The SLE configuration (part 3) : This part consists of three tables • The land use configuration table : the user can edit the geo-typical configuration (texture and vegetation) for each land-use class. • The object importer table : the user can import a 3D object and insert it at a specific place. • The rendering table : the user can edit rendering options.
4.3.2 The land use configuration
When a project is loaded into SLE, you need to configure the texture and 3D objects to add for each land use type.
Page 26 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Then, the user has to parameterize the configuration table.
Each row of this table represents a land use. For each land-use, the user can edit geo-typical parameters (in columns): • The texture represents typical ground cover. This is used to build the texture that is draped over the terrain when rendering. Double-click on a cell of the Texture column to open a file browser and choose an image file (bmp, png, or jpg format). • The resolution of the texture. Click on a cell of the Resolution column and enter a value. This value is the size in metres of a texture pixel. • The density of 3D objects on this land use. Click on a cell of the Density column and enter the average number of objects by m². • The list of 3D objects on this land use. When you double-click on a cell of the Mesh column, a new window with the following parameters appears:
Page 27 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
In this panel, the user can add meshes to the selected land-use by clicking on Add Mesh . A new line in the table is created where the user can edit: o The 3D object file. It is possible also to define objects like rocks, as long as they have a natural distribution: double-click on a cell of the Mesh column to open a file browser and choose a mesh file (.mesh format only). o The type of spatialization: the object’s distribution over the terrain can be done randomly or by rows. Double-click on a cell of the Spatial column to enter rand (for random positions) or line ( for aligned positions). o The density. This is used in conjunction with a random offset to determine the number of objects in a specific area. Double-click on a cell of the Density column to enter a value between 0 and 100. o The inter-row/intra-row spacing. This is used in conjunction with row spatialization to specify the distance between objects. Double-click on a cell of the Inter-row column or Intra-row column to enter the spacing value in metres. o The scale variation information. This is used to generate variations in the appearance of objects, and is particularly useful when billboards are used to render trees. Double-click on a cell of the Scale column to enter a value between 0 and 100. o The level of detail. The object can be displayed in 2D (billboards) or in 3D according the distance of the camera. Click on the checkbox of the Full Geometry column The user can remove a mesh by clicking on Remove Mesh button. Then he needs to save the mesh configurations for this land use. Returning to the previous window, he just has to click on “compute terrain” to start the visualization.
4.3.3 The 3D visualisation
According to the amount of data, the user needs to wait before the 3D display. Once computed, the 3D landscape is loaded and the user can move into the scene with the keyboard (see above for the control keys). A compass can help the user for the navigation.
Page 28 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
4.3.4 The 3D object importer
- Click on Browse Lig: Open a File Browser. The user can choose AMAP Lig file format. - Click on Browse Mesh: Open a File Browser. The user can choose Mesh file format. Once selected, the 3D object is loaded in the 3D view. The user can move the object with Right Click + Mouse Cursor Right. On Left Click the object is positioned.
4.3.5 The rendering option
1
2 3
4 5
6
7
8
9
10 11
12
13 14
15
16 17
18 19 20
Page 29 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
1. Shot: Take a screenshot. Save the current viewport on the current directory (SLEInstallDirectory\bin\release \screenshot_ i.png where i is the screenshot number). 2. Full 3D: When activated, all the geometry in a 200 m diameter circle around the camera are displayed when the user takes a screenshot. 3. Shadows: When activated, shadows in a 200 m diameter circle around the camera are displayed when the user takes a screenshot. 4. Render Sky: When activated, the sky is rendered. The user can parameterize the sky rendering with the options below. 5. Modify Sun Position: Slider to modify the sun time. The centre slider position is midday. 6. Time Speed: Slider to modify the time speed. By default, the time is paused. When the slider is moved to the right, the time speed is increased. 7. Modify cloud density: Slider to modify the cloud density. When the slider is moved to the right, cloud density is increased. 8. Modify fog density: Slider to modify the fog density. When the slider is moved to the right, fog density is increased. 9. Modify rain density: Slider to modify the rain density. When the slider is moved to the right, rain density is increased. Caution: the time speed must be >0. 10. Lighting Power: When activated, the HDR (Hight Dynamic Range) rendering is activated. http://en.wikipedia.org/wiki/High_dynamic_range_rendering 11. HDR Velocity: Slider to modify the HDR velocity. When the slider is moved to the right, light velocity is increased. 12. Compute: Compute the terrain light map, including the terrain shadows. 13. Render Stereo: When activated, the stereo-rendering is enabled. 14. Camera Distance: Slider to modify the distance between the two cameras. When the slider is moved to the right, the distance is increased. 15. Focal Distance: Slider to modify the focal. When the slider is moved to the right, focal distance is increased. 16. Render Hydro: When activated, a water plane is added to the scene. 17. Z Position: Slider to modify the elevation of the water plane. 18. X Position: Slider to modify the abscissa position of the water plane. 19. Y Position: Slider to modify the ordinate position of the water plane. 20. Size of the plane: Slider to modify the area of the water plane. When the slider is moved to the right, the plane size is increased.
Page 30 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
4.3.6 Recording & playing back camera paths
You can start and stop recording the camera by pressing Insert , and toggle playback using Return . You can also place keyframes manually with the Space key. Positions between keyframes are interpolated with splines. You can use the brackets keys [ and ] to change the time interval between the keyframes that are created in record mode, and the playback speed in playback mode. Once the playback is started, frames are saved on the current directory (SLEInstallDirectory\bin\release ) in PNG format ( movie i.png where i is the frame number). When playback is finished, makemovie.bat creates automatically the movie file ( movie.mp4 ).
Page 31 of 46
SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
5 Illustrative application
In this chapter, we present an example of SLE visualisation for a landscape in the South of France, resulting from different scenarios (Nespoulous, 2004).
5.1 The situation
For centuries, the North of the Mediterranean basin has been characterized by complex and quick changes. These changes mark a break with the progressive variations which took place since the Neolithic and contributed to manufacture the landscape in mosaic which is considered as typical for the North of the Mediterranean basin. The nature of these changes affecting biodiversity of the Mediterranean basin raises the question of the impact of human activities on biodiversity, which is a central stake in the development of territories (Cheylan and Gumuchian 2002). The studied zone is the Pic Saint Loup region, located in the department of Hérault at about twenty kilometres to the Northwest of Montpellier. This territory is characterised by the mountain of the Pic Saint Loup reaching the altitude of 658 metres. The natural vegetation is dominated by "garrigue", the typical Mediterranean shrubland ecosystem characterised by discontinuous bushy associations of calcareous plateaus, often composed of kermes oak, lavender, thyme, and white cistus, with isolated trees (holm oak and juniper). Agriculture is dominated by pasture, mainly in silvopastoral systems, olive groves and vineyards, with rare agricultural crops (wheat and barley). Figure 9: Localization of the Pic Saint Loup area (France)
Page 33 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
5.2 The scenarios
A detailed study of the past and present land-use on the territory of Pic Saint Loup, based on aerial photography in years 1946, 1962, 1981, 1992 and 2002, showed an important impact of peri-urban development on biodiversity (Sirami et al., 2007, 2008). Following these results, four scenarios were set up, as part of a participatory process for planning the regional peri- urban and agricultural policy. They can be summarized according to two axes (Figure 10). The abscissa concerns "nature", which is more or less rich in biodiversity, corresponding to the environmental quality of natural areas, and the ordinate fluctuates between urban and rural, which determines the type of urbanization and way of life favoured in each scenario (Nespoulous, 2004). Land-use maps were produced for each scenario and the resulting scenes were processed with SLE. Two representative viewpoints were chosen, one general view of the landscape, and one characteristic view dominated by the Pic Saint Loup mountain. Figure 10: Four scenarios of the Pic Saint Loup area in 2030
Nature +++
Urban Rural
34 Nature
---
Page 34 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
5.2.1 Scenario 1: Biodiversity by agriculture
This first scenario results from the realization of the benefit of quality agriculture, in the biodiversity and Mediterranean landscape conservation. This realization is supported by the collapse of the international transport, which implicates a relocation of agricultural production. In this context, the balance between the housing settlement and the workplace is proclaimed as a major target in town and country planning. Therefore, the village growth is self-restraint and new house settlements are made closed to the existent habitat, to use less space. The traditional agrarian activities, of picking, viticulture and olive-growing reinvest garrigues. The forest is grazed and cut for wood energy production. As a consequence, the landscape is opened by agrarian activities, but the practices diversity allows the preservation of an environmental mosaic. Figure 11: Biodiversity by agriculture
Page 35 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
5.2.2 Scenario 2: A green city in a Mediterranean forest
In this second scenario, the development of the big cities, as Montpellier, is made in further villages. However, public policies structure this urban growth and are very sensitive to the natural environment quality. Therefore, the village extension continues, close to historical cores and along the main axes of circulation, in keeping with the nature protection, according to environmental models of landscaping. On the contrary, agriculture does not belong anymore really to the garrigues, it remains some relics of high quality historical grapevines. Landscape are then very wooded, since the emphasis was the Mediterranean forest. Figure 12: A green city in Mediterranean forest
Page 36 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
5.2.3 Scenario 3: Urban Pressure
For this third scenario, the urban development is more important and this growth is not supervised by land-use planning and development. The urban model growth is along the road axes and close to the urban centres and is stronger than for the second scenario. In this context, it is important to develop the axes of circulation between the main city and its periphery. Agriculture is also present and is based on two different models. On the one hand, there are sustainable yields, based on the vineyards AOC reputation, and on the other hand there is subsistence agriculture around settlements. Figure 13: Urban pressure
Page 37 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
5.2.4 Scenario 4: The garrigue after energy crisis
With the collapse of the oil system, populations have to find alternative energy. A direct consequence of this crisis is the increase of transport costs and therefore requires re- concentrating the productions at local level. Thus, the life in garrigues folds up on itself and leads to the production of energy but also subsistence crops. The land is then broadly agrarian and opened, composed of big open fields parcels, to acquire a better output. On hills, landscape take a forested turn of stamp post, with very big plots of cutting areas. In the south, fields of solar panels are installed, and aeolian on the blowy summits, All natural resources are used to produce energy. As this scenario is a crisis period, urbanization is decelerated. Figure 14: The garrigue after energy crisis
Page 38 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
6 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 for example Appleton and Lovett 2005, Bishop et al., 2005, Bishop et al., 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 for agri-environmental policies, such as the one developed in the SEAMLESS project (Van Ittersum et al., 2008). The 'SLE' module has not up to now been integrated in the 'SEAMLESS Integrated Framework', but this possibility is planned for future developments. 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 precise 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. Disaggregating the results remains a difficult task. Le Ber et al., (2006a) 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 and 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 the chapter on Frameworking with indicators: contributions to a systemic approach, by Geniaux et al., and D2.1.3). 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 participative planning. In the BioScene project Soliva and Hunziker (2008) coupled ideal type narratives with computer-aided photo-editing for participatory landscape planning. In Australia, Bishop et al., (2008) propose agent-based modelling to explore the decision-making process, in a "virtual decision environment". In SLE it is possible to visualize an additional GIS layer : for example, in the urban pressure scenario, we computed 3D views with an abstract indicator representing probable urban expansion area (Figure 15). Indeed, the introduction of new large structures such as a city Page 39 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009 represents visual intrusions which may reduce visual quality. This quality decrease is related to the level of modification such as contrast in size, shape, colour and texture between the structure and the pre-existing landscape. The magnitude of the impact can be considered to depend on the extent of the area affected (Rivas et al., 1997) and indicators of magnitude can be for example the total area from which the new structure can be seen, as in Figure 15. It is also interesting to use GIS maps to display non-visual information such as wildlife migration corridors, water flow, remoteness, etc. Figure 15: Spatial patterns of urban land use growth (shaded in purple)
There has been much discussion concerning the ethics of using landscape visualization, and several different issues can be raised. One of them concerns the relation between ecology or ecosystem quality and aesthetics: Gobster et al., (2007) argue for instance that future landscape patterns, human experiences, and actions can be devised to create landscapes that are ecologically beneficial and simultaneously elicit aesthetic pleasure. However, the aesthetics of a virtual landscape can be influenced by the degree of realism of the visualization tools (Lange 2001, Daniel 2001, MacFarlane et al., 2005), and it is necessary to define what is the "sufficient realism" for environmental decision making (Appleton and Lovett 2003) . In addition, using virtual imagery can lead to a strong bias due to the selection or highlighting of particular aspects in order to persuade the public on particular environmental issues (Appleton and Lovett 2005, Sheppard 2005, 2006). Sheppard and Cizek (2008) show several examples of misuse of landscape visualization, and suggest a code of ethics, with "the combination of scientific/technical expertise, 3D computer modelling skills, and understanding of social responses to landscape imagery".
Page 40 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Conclusion
Decisions support systems are increasingly being applied in spatial planning, and virtual landscapes become an important part of decision making. Planners recognise realism as an important factor in this type of visualisation (Appleton and Lovett, 2005). It is important to define an appropriate level of realism, because the “photorealism” can have potential negative effects if it is not linked to real-world data. Furthermore, the very fact that we have so much control over the content and style of a visualisation means that everything must be questioned – viewpoint choice, presentation method and addition of auxiliary information should all be considered alongside realism issues when creating images for planning purposes, since they are not subject to the limitations imposed by photo-based or artistic techniques, and they all have the potential to affect the feedback gained from a consultation exercise. The opportunities presented by advancing technology should not be automatically taken, but carefully evaluated and implemented with regard to the needs of the project in question. Only then will computer-generated visualisations form a useful and reliable part of the planning process.
Page 41 of 46
SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
References
Appleton, K. & Lovett, A. (2003). GIS-based visualisation of rural landscapes: defining ‘sufficient’ realism for environmental decision-making. Landscape and Urban Planning, 65 , 117-131. Appleton, K. & Lovett, A. (2005). GIS-based visualisation of development proposals: reactions from planning and related professionals. Computers, Environment and Urban Systems, 29 (3), 321-339. Appleton, K., Lovett, A., Sünnenberg, G. & Dockerty, T. (2002). Rural landscape visualisation from FIS databases: a comparison of approaches, options and problems. Computers, Environment and Urban Systems, 26 , 141-162. Auclair D., Barczi J.-F., Borne F. and Étienne M. (2001a). Assessing the visual impact of agroforestry management with landscape design software, Landscape Research 26 (2001), pp. 397–406. Auclair, D., Barczi, J.F., Borne, F. & Etienne, M. (2001b). Landscape visualisation software as a forest management decision support system. In A. Franc, O. Laroussinie & T. Karjalainen (Eds.). Criteria and indicators for sustainable forest management at the forest management unit level . Torikatu: EFI, p. 207-214. Bergen, S.D., Fridley, J.L., Ganter, M.A. and Schiess, P. (1995). Predicting the visual effect of forest operations. Journal of Forestry, 93: 33-37. Bell, S. (2001). Landscape pattern, perception and visualisation in the visual management of forests. Landscape and Urban Planning, 54 , 204-211. Bishop I.D, Wherrett J.R. and Miller D.R. (2001). Assessment of path choices on a country walk using a virtual environment, Landscape and Urban Planning 52 (2001), pp. 225– 237. Bishop I.D., Hull R.B. and Stock C. (2003). Envisioning Systems: software with a world view, 5th International Symposium on Environmental Software Systems Semmering Austria. Bishop I.D., Hull R.B. and Stock C. (2005). Supporting personal world-views in an envisioning system, Environ. Model. Software 20, pp. 1459–1468. Bishop, I.D., Stock, C. & Williams, K.J. (2008). Using virtual environments and agent models in multi-criteria decision-making. Land Use Policy, 26, 87-94. Bloom, C. (2000). Terrain Texture Composition by Blending in the Frame Buffer ("Splatting Textures"). www.cbloom.com/3d/techdocs/splatting.txt . Boudon, F., Nouguier, C. and Godin, C. (2001). GEOM module manual. II. Developer guide. Document de travail du Programme Modélisation des plantes, DT4-2001. CIRAD, Montpellier, France. 600 p. Cartwright W. (2008). Visualising Alternative Futures. In C. Pettit, W. Cartwright, I. Bishop, K. Lowell, D. Pullar & D. Duncan (Eds.), Landscape analysis and visualisation: spatial models for natural resource management and planning (pp. 490-507). Berlin: Springer. Cheylan J.P., Gumuchian H. (2002). L’évaluation des impacts des politiques publiques paysagères au sein des Parcs Naturels Régionaux en montagnes méditerranéennes: quelles méthodes? Quels outils?, Méthodes et outils pour l’évaluation des impacts des PPP, Numéro spécial: Politiques publiques paysagères et Parcs Naturels Régionaux; pour une évaluation, pp 13-24.
Page 43 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Danahy J.W. (1989). Irises in a landscape: an experiment in dynamic interaction and teaching design studio. In: M. McCullough, W.J. Mitchell and P. Purcell, Editors, Proceedings of the CAAD Futures’89, The electronic design studio: Architectural knowledge and media in the computer era, MIT Press, Cambridge, MA (1989), pp. 363–376. Daniel, T.C (1992). Data visualization for decision support in environmental management. Landscape Urban Planning, 21 , 261–263. Daniel, T.C (2001). Wither scenic beauty? Visual landscape quality assessment in the 21 st century. Landscape Urban Planning, 54 , 267–281. Day, A.M. & Willmott, J. (2005). Compound textures for dynamic impostor rendering. Computers & Graphics , 29 : 109–124. De Boer, W.H. (2000). Fast Terrain Rendering Using Geometrical Mipmapping. http://www.flipcode.com/articles/article_geomipmaps.shtml . Dockerty, T., Lovett, A., Sünnenberg, G., Appleton, K. & Parry, M. (2005). Visualising the potential impacts of climate change on rural landscapes. Computers, Environment and Urban Systems, 29 , 297-320. Ervin, S.M. & Hasbrouck, H.H. (2001). Landscape modelling: digital techniques for landscape visualization . New York: McGraw-Hill. Gardner, R.H. & Urban, D.L. (2006). Neutral models for testing landscape hypotheses. Landscape Ecology, 22 , 15-29. Ghadirian, P. & Bishop, I.D. (2008). Integration of augmented reality and GIS: a new approach to realistic landscape visualisation. Landscape and Urban Planning, 86 , 226- 232. Gobster, P.H., Nassauer, J.I., Daniel, T.C. & Fry, G. (2007). The shared landscape: what does aesthetics have to do with ecology? Landscape Ecology, 22 , 959–972. Google Inc. (2006). SketchUp User's Guide. 362 p. http://download.sketchup.com/GSU/pdfs/GSUUsersGuide_WIN.pdf . Griffon, S. & Auclair, D. (2009) in preparation. Helming, K., Tscherning, K., König, B., Sieber, S., Wiggering, H., Kuhlman, T., Wascher, D., Perez-Soba, M., Smeets, P., Tabbush, P., Dilly, O., Hüttl, R. & Bach, H. (2008). Ex ante impact assessment of land use changes in European regions — the SENSOR approach. In K. Helming, M. Perez-Soba & P.Tabbush, Sustainability Impact Assessment of Land Use Changes . Berlin: Springer. Herwig A. and Paar P. (2002). Game engines: tools for landscape visualization and planning? In: E. Buhmann, U. Nothelfer and M. Pietsch, Editors, Proceedings at Anhalt University of Applied Sciences, Trends in GIS and virtualization in environmental planning and design, Wichmann, Heidelberg (2002), pp. 162–171. Lange, E. (2001). The limits of realism: perceptions of virtual landscapes. Landscape and Urban Planning, 54, 163–182. Lange, E. & Bishop, I. (2001). Our visual landscape: analysis, modelling, visualization and protection. Landscape and Urban Planning, 54 , 1-3. Lange, E., Hehl-Lange, S. & Brewer M.J. (2008). Scenario-visualization for the assessment of perceived green space qualities at the urban-rural fringe. Journal of Environmental Management, 89 , 245-256. Le Ber, F., Benoît, M., Schott, C., Mari, J.-F. & Mignolet C. (2006a). Studying crop sequences with CarrotAge, a HMM-based data mining software. Ecological Modelling, 191, 170-185.
Page 44 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Le Ber, F., Lavigne, C., Mari, J.F., Adamczyk, K. & Angevin, F. (2006b). GENEXP, un logiciel pour simuler des paysages agricoles en vue de l'étude de la diffusion de transgènes . (Paper presented at the SAGEO 2006 International Conference on Spatial Analysis and GEOmatics; Research & Development, Strasbourg, FR. Retrieved from http://hal.archives-ouvertes.fr/docs/00/10/33/82/PDF/sageo06.pdf . Lovett, A. (2005). Futurescapes. Computers, Environment and Urban Systems, 29, 249-253. Lovett A., Kennaway,R. Sünnenberg G., Cobb D., Dolman P. and O’Riordan T. (2001). Visualising sustainable agricultural landscapes. In: D. Unwin and P. Fisher, Editors, Virtual reality in geography, Taylor & Francis, London (2001), pp. 102–130. MacFarlane, R., Stagg, H., Turner, K. & Lievesley, M. (2005). Peering through the smoke? Tensions in landscape visualisation. Computers, Environment and Urban Systems, 29 , 341-359. Mansergh, I., Lau, A. & Anderson, R. (2008). Geographic Landscape Visualisation in Planning Adaptation to Climate Change in Victoria, Australia. In C. Pettit, W. Cartwright, I. Bishop, K. Lowell, D. Pullar & D. Duncan (Eds.), Landscape analysis and visualisation: spatial models for natural resource management and planning (pp. 469-487). Berlin: Springer. Muhar, A. (2001). Three-dimensional modelling and visualisation of vegetation for landscape simulation. Landscape and Urban Planning, 54 , 5-17. Nespoulous A. (2004). Relations entre la dynamique de la végétation et la gestion sociale de l’espace: les garrigues du Pic Saint Loup, MA Thesis, Univ. Montpellier III, 89p. Orland B., Budthimedhee K. and Uusitalo J. (2001). Considering virtual worlds as representations of landscape realities and as tools for landscape planning, Landscape and Urban Planning 54, pp. 139–148. Paar, P. & Clasen, M. (2007). Earth, landscape, biotope, plant. Interactive visualisation with Biosphere3D. In M. Schrenk, V.V. Popovich & J. Benedikt. Real Corp 007 - To plan is not enough: strategies, plans, concepts, projects and their successful implementation in urban, regional and real estate development. Vienna: CORP. Perrin, L., Beauvais, N. & Puppo M. (2001). Procedural landscape modelling with geographic information: the IMAGIS approach. Landscape and Urban Planning, 54 , 33-47. Pettit, C., Cartwright, W., Bishop, I., Lowell, K., Pullar, D. & Duncan, D. (Eds.) (2008). Landscape analysis and visualisation: spatial models for natural resource management and planning. Book series: Lecture notes in geoinformation and cartography. Berlin: Springer. Pradal C., Donès N., Godin C., Barbier de Reuille P., Boudon F., Adam B., Sinoquet H. (2004). ALEA: A software for integrating analysis and simulation tools for 3D architecture and ecophysiology. In: proceedings of the 4th International Workshop on Functional-Structural Plant Models, Godin C. et al. (Eds), Montpellier, France 7-11 june 2004. Montpellier. pp. 406. Rivas V, Rix K., Frances E., Cendrero A., Brunsden D (1997). Geomorphological indicators for environmental impact assessment; consumable and non consumable geomorphological resources. Geomorphology 18 (3-4), 169.182. Salter, J.D., Campbell, C., Journeay, M. & Sheppard, S.R.J. (2008). The digital workshop: Exploring the use of interactive and immersive visualisation tools in participatory planning. Journal of Environmental Management (in press) , doi:10.1016/j.jenvman.2007.08.023.
Page 45 of 46 SEAMLESS No. 010036 Deliverable number: D3.7.8.2 18 January 2009
Sauget, N. and Depuy, M. (1996). Forêt paysanne et paysage: les agriculteurs et le visible. In: G. Balent (Editor), La forêt paysanne dans l'espace rural. Biodiversité, paysages, produits. Étud. Rech. Syst. Agraires Dév., 29, pp. 245-264. Savill, P., Evans, J., Auclair, D. and Falck, J. (1997). Plantation silviculture in Europe. Oxford University Press, Oxford, UK, 297 p. Sheppard, S.R.J. (1982). Landscape portrayals: Their use, accuracy, and validity in simulating proposed landscape change. Dissertation, University of California, Berkeley. Sheppard, S.R.J. (2005). Landscape visualisation and climate change: the potential for influencing perceptions and behaviour. Environmental Science and Policy, 8, 637-654. Sheppard, S.R.J. (2006). Bridging the sustainability gap with landscape visualisation in community visioning hubs. IAJ the Integrated Assessment Journal, 6(4), 79-108. Sheppard, S.R.J. & Cizek, P. (2008). The ethics of Google Earth: Crossing thresholds from spatial data to landscape visualisation. Journal of Environmental Management, (in press). doi:10.1016/j.jenvman.2007.09.012. Sirami, C., Brotons, L. & Martin J.L. (2007). Vegetation and songbird response to land abandonment: from landscape to census plot. Diversity and Distributions, 13, 42–52. Sirami, C., Brotons, L. & Martin J.L. (2008). Spatial extent of bird species response to landscape changes: colonisation/extinction dynamics at the community-level in two contrasting habitats. Ecography, 31 , 509-518. Snyder K. (2003). Tools for community design and decision making. In: S. Geertman and J. Stilwell, Editors, Planning support systems in practice, Springer, Berlin (2003), pp. 99–120. Soliva, R. & Hunziker, M. (2008). Beyond the visual dimension: Using ideal type narratives to analyse people’s assessments of landscape scenarios. Land Use Policy (in press ), doi:10.1016/j.landusepol.2008.03.007. Thomas, A.L. and Price, C. (1999). Landscape valuation of farm woodlands. In: P.J. Burgess, E.D.R. Brierley, J. Morris and J. Evans (Editors), Farm woodlands for the future. BIOS Scientific Publishers, Oxford, UK, pp. 69-79. Tress, B. & Tress, G. (2003). Scenario visualisation for participatory landscape planning — a study from Denmark. Landscape and Urban Planning 64 , 161–178. Tyrväinen, L. and Tahvanainen, L. (2000). Landscape visualisation in rural land-use planning. In: XXI IUFRO world congress. Forests and society: the role of research, Kuala Lumpur, Malaysia, vol. 1, pp. 338-347. Van Ittersum, M.K., Ewert, F., Heckelei, T., Wery, J., Alkan Olsson, J., Andersen, E., Bezlepkina, I., Brouwer, F., Donatelli, M., Flichman, G., Olsson, L., Rizzoli, A.E., Van der Wal, T., Wien, J.E. & Wolf, J. (2008). Integrated assessment of agricultural systems – A component-based framework for the European Union (SEAMLESS). Agricultural Systems, 96 , 150-165. Wissen, U., Schroth, O., Lange, E. & Schmid, W.A. (2008). Approaches to integrating indicators into 3D landscape visualisations and their benefits for participative planning situations. Journal of Environmental Management, 89, 184-196.
Page 46 of 46