Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 583

Connection between forest inventory data and geographic information systems for assessing timber value at the stand level

F. BLAISE 1, L. SAINT-ANDRÉ 2, J.M. LEBAN 3, J.C. GÉGOUT 3, J.C. HERVÉ 3.

1 Loria-INRIA, Technopôle de Nancy-Brabois, Campus scientifique, 615 rue du Jardin Botanique, B.P. 101, 54602 Villers-les-Nancy CEDEX (France)

2 CIRAD-forêt, Campus international de Baillarguet – TA 10 / D,

34398 Montpellier CEDEX 5 (France)

3 LERFoB, joint ENGREF-INRA Research Unit “Forest and Wood Resources”, INRA, Equipe de recherches sur la qualité des bois, Champenoux, 54280 Seichamps (France)

INTRODUCTION Recent developments in the modelling of tree growth and wood quality have produced new models and software packages that enable the simulation of timber properties, usually at the stand level. Separately foresters are now deeply involved in the development and implementation of GIS concerning forest yield, vegetation, soil properties, climatic variables etc. To our knowledge there is no wood quality information available in such a data base. The objective of this work is to demonstrate that it is possible to establish a linkage between these Growth and Wood Quality models and the available forest geographic information. We intend to present a method which enables to produce for the French Vosges mountain area (i) a map of forest ecological conditions (which combine climatic, trophic and hydric characteristics of sites), (ii) a map of the sawn timber properties for Norway spruce stands and (iii) a realistic 3D representation of trees for one or more selected stands. The software involved for this work are (i) AMAP that alows to simulate and to stage 3D shapes of plants including trees species , (ii) ARCINFO© a commercial GIS, (iii) IMAGIS which interfaces AMAP and GIS in order to represent GIS information in a realistic way, (iv) Win-EPIFN which simulates the board properties from one stand described by usual forest inventory measurements. GROWTH MODELS: DIFFERENT APPROACHES Forests perform multiple functions and constitute sensitive ecosystems which require not only to be protected, but also to be highlighted and durably managed. In this purpose, the processes of growth and competition which are responsible for the dynamics and the evolution of the plantings, regular or irregular, are analysed, according to the forestry. Models of growth and production are established and integrated into decision-tools for the management. There are several types of classic approach in modelling of growth and fruit production: • The models of statistical type, based on the techniques of regression. They can concern (i) the evaluation of the

average behaviour of the plantings and the forest covers (stand models, eg. Maugé 1975), (ii) the evaluation of individual tree growth depending on the stand density and the tree social status (non-spatially explicit tree models or distance-independent models, eg. Dreyfus 1993, Dhôte 1995, Meredieu 1998, Saint-André et al. 2002), (iii) the evaluation of individual tree growth depending on the tree position within the stand (spatially explicit tree models or distance-dependent models, eg. Gourlet-Fleury 1999, Gourlet-Fleury and Houllier 2000, Courbaud et al. 2001) (Fig.1).

• The models of fruit production in controlled environment integrate physical parameters (temperature, light,

resources) and knowledge in physiology (processed-based models, e.g. Mäkelä 1999, Loustau et al. 2001). They are especially used on planted crops (tomatoes, cotton plant) and are currently widely developed for forest stands. The morphology of the plant is there unknown or roughly considered, although produced biomass is calculated (Fig.2).

• The models of competition. At these last ones, rudimentary morphological parameters are taken into account. In the

case of forest applications, the simple models of crown are implemented and allow the estimation of the wooden

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 584

production according to the stand density. Interactions between trees are administered by contacts between simple engendered volumes (Fig.3).

• The morphological models which build 3D models by computer simulation. They are based on botanical knowledge in vegetable architecture, and are mainly descriptive. The vegetable architecture possesses indeed its own rules of internal organization, although the result of a growth is connected to the physical parameters of the environment. The same vegetable structure can have been realized more or less quickly under different growing conditions. Most of existing morphological models result from the computer science and implement algorithms specialized in the construction of ramified structure. Qualitative knowledge integrated by these models are rather summary and their use is situated essentially in the field of the image synthesis (Fig.4-5).

Other specialized types of model are used in mechanics or in transfer of elements.

Fig. 1 : CAPSIS software (CIRAD-AMIS, 2000). Fig. 2 : Simulation of tomato plant (Gary et al., 1994).

Fig. 3 : The program Pinogram (Pine GRowth Area Model): simulation of pine plantation with competition (Leersnijder, 1997).

CO2 enrichment Reference First truss pruned

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 585

Fig. 4 : Simulation of two conifers with competition Fig. 5 : Spruce simulation (Lintermann and Deussen, 2000). (Prusinkiewicz et al., 1994). Ultimately, all these models are complementary the ones to the others and, obviously, a general model which would include at the same time the problems of morphology and interaction with the environment would have a remarkable versatility in its agronomic applications. WORKS OF THE LERBOB IN FOREST RESEARCH General presentation The models of forest production aim to simulate on the long term, and with a good reliability required by the forest administrative customers, the evolution of the forests on the scale of the plot (10 hectares). They take as inputs the ecological characterization of the environment (fertility of the station, total changes a long the rotation) and the actions of forest managers throughout the life of the plot. They provide as outputs statistics summarizing the state of the plot (number of trees, diameters, heights and volumes). These statistical exits are well adapted to the forester needs, but they remain rather abstract if the goal is to communicate or negotiate with "profane" interlocutors. In this case, taking into account the visual aspect (of a plot, of a path in forest) in management projects would be an invaluable tool. Within this framework, LERFoB (Laboratory of study of the Resources Forest-Wood, UMR INRA-ENGREF) develops a whole of quantitative methods to modernize forest management: • NTM (Numerical Terrain Model), GIS (Geographic Information System) and forest ecology. The goal is to create

methods of spatialized treatment for the ecological determinants of the dynamics of forests (relief, climate, grounds, vegetation), by interfacing data bases (NTM and ecological data bases) and simulators (Fig.6).

• Dynamic systems and modelling of forest dynamics: the goal is to develop robust models and simulators to predict the long term evolution of the forests, in an ecological and socio-economic changing environment (software Fagacées, CAPSIS).

• Simulation of the quality of the forest products (Fig.7): the goal is to build methods to evaluate the wood properties of a given standing forest resource (WinEPIFN, software of resource evaluation).

Fig. 6 : Non homogeneous environment: there are differences of hillsides (left picture); these oppositions lead to differences of radiative intensity according to hillsides (right picture).

Fig. 7 : Not destructive estimate of the quality of a log: 3D reconstruction from scanned views.

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 586

WinEPIFN software WinEPIFN® is a software which aims at predicting the quality of logs and of sawn boards that would result from the harvesting of standing forest resources (Leban et al., 1996). The software can be used at the stand level (e.g. for assessing the value of a compartment to be sold or purchased), at the planning unit level (e.g. for assessing the average quality within a given unit made of several stands), at the regional level (e.g. within the framework of a national forest inventory). The underlying method was first developed and validated for Norway spruce in Northeastern France (Houllier et al., 1995). It has been then extended to other French regions for the same species, and to Corsican pine and Douglas fir. It can be easily adapted to any coniferous species which has a straight stem, a symmetrical crown development and a strong apical dominance. The method is based on a chain of empirical tree growth and branch models that aim at predicting the inner stem structure of a given tree (i.e., branches, knots and ring distribution) from simple field measurements; this chain of models is completed by geometrical models and grading rules that convert the inner stem structure into graded logs and/or graded boards. Input data consist of a list of trees that may belong to a single stand or to several distinct stands (e.g. all the sample plots of the national forest inventory in a region) and for which usual inventory data are available: age, diameter at breast height, total height. These input data are completed by a log bucking and a board sawing pattern. As an alternative, WinEPIFN® can also be fed with lists of trees generated by a growth and yield model (e.g. by a model implemented as a CAPSIS® module). According to the resolution of that model (e.g. whether stem inner structure is predicted or not), some of the modules of WinEPIFN® may be skipped. Output data consist of: (1) the prediction of the inner structure of all trees listed in input data; (2) the predicted description of all the logs and their grade (if a log grading rule is available); (3) the predicted description of all the boards and their grade (if a board grading rule is available) (Fig.8).

Fig. 8 : WinEPIFN® outputs.

Future developments will consist in adapting this approach to other coniferous species (e.g. Atlas cedar, maritime pine) but also to broadleaf species (e.g. oak, beech, eucalypt and teak) and to mixed and uneven-aged stands. This will raise new problems (e.g. stem shape, reaction wood, forks, low large branches). A similar approach can also be used to assess the mineral contents (e.g., in carbon, nitrogen or phosphorus) in a tree and, by extension, in a stand: this was already tested on douglas fir. AMAP AND THE TREE ARCHITECTURE MODELLING Here, we are interested in the architecture of the plants and with the processes which ensure its operation. This approach is privileged in the work of the Program of Plant Modelling of CIRAD (laboratory AMAP). It is based on the qualitative knowledge brought by the school of Hallé and Oldeman in architecture and on the quantitative methods developed at the point with AMAP. These methods are based on the description of the buds dynamics (growth, death, ramification) by stochastic processes (Blaise et al., 1996 ; Blaise et al., 1998 ; Reffye (de) and Houllier, 1997). The simulation software resulting from these theories makes it possible to build realistic models of plants in which topology

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 587

(relationship between the entities constituting the plant) (Fig.9.a) and the geometry (size of these entities) (Fig.9.b) are simulated according to the parameters of the model, themselves estimated from the experimental data catches on the plants. These models allows to visualize in 3D the architecture of the plants (Fig.10.a) or the forest covers (Fig.10.b), but can also be used as a basis for particular applications. They make it possible for example to simulate the penetrating solar radiation in a vegetable cover and to calculate the radiative transfers (Fig.11).

Fig. 9 : Topological structure (a) and geometrical structure (b).

Fig. 10 : Plant growth (a) and forest (b) simulated with AMAP.

Fig. 11 : Simulated map of transmitted PAR (Photosynthetic Active Radiation) under a 20-year-old coconut plantation (Dauzat and Eroy, 1996).

FROM GIS DATA TO 3D: IMAGIS SOFTWARE Presentation

(a) (b)

(a) (b)

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 588

IMAGIS (IMAGIS, 1996) is the implementation result of the European project Imago Metropolis. This project of transfer and technological validation launched in 1996 was supported by the European Commission within the framework of the Program INNOVATION. The IMAGIS software aims to represent, in the manner the realistic photos, the information contained in a GIS (Fig.12). The user has an interactive tool allowing him to extract the information contained in the GIS data base, to constitute and handle a three-dimensional scene including/understanding of the 3D forms representative of the contents of the various objects of the data base, to choose a point of view of this scene and to calculate the image corresponding. The combination of IMAGIS with AMAP can be used to semi-automatically generate 3D scenes from 2D data contained in GIS. Moreover, the representation of great landscapes underlies the concept of multiplicity of forms and thus an enormous quantity of polygons. To obtain reasonable computing times for images, it is then necessary to develop an engine making it possible to automatically simplify the models of three-dimensional forms according to the distance to which they are seen, without decreasing too much the quality of the images thus calculated.

Fig. 12 : IMAGIS structure: 3D representation of a GIS. General functioning IMAGIS is an interface between the worlds of cartography and image synthesis. It is known to offer to the users of GIS a toolbox enabling them to transform occupation maps of the grounds into models of landscape in 3D. This progressive transformation is operated by specific procedures where very synthetic 2D information of the GIS is enriched by 3D information (Auclair et al., 2000 ; Barczi et al., 2000). To carry out this transformation, the user establishes links between the entities coming from the GIS (the map) and one or more generic objects of the IMAGIS base. Each of them manages a particular procedure making it possible to have of the 3D forms and/or textures on a NTM, to generate envelopes of buildings… In a GIS structure, every entity of the landscape has a specific attribute (the class) in the naming. When a link is established between a class of the GIS and one of IMAGIS's generic objects, a procedure of specific modelling is initialised. It uses as input the geometry of the elements of the GIS as well as the values assigned to the parameters of the generic object. It produces as output groups of 3D forms sharing the same morphological characteristics. FIRST RESULTS: APPLICATION TO THE VOSGES MOUNTAINS At first, the Program of Plant Modelling of the CIRAD developed a range of specific software, AMAP INTEGRAL™, for the conception and the visualization of landscapes. This range consists of tools or modular workshops allowing the simulation of the growth and the visualization of plants, as well as the conception and the simulation in the space and the time of projects of organization. These software are at present marketed by the company Bionatics ® (Bionatics, 2000). So the various stages to represent a forest cover in 3D can be summarized: (i) terrain modelling, either from GIS data or from file of 3D points, (ii) IMAGIS to make the link between GIS data and IMAGIS’s generic objects, (iii) AMAP INTEGRAL™ to perform the 3D archictecture of the trees and (iv) AMAP INTEGRAL™ for the scene visualisation.

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 589

Terrain modelling The Terrain workshop is a numeric modeller of terrains, for creating a topography or of reconstituting it from a file of 3D points (Fig.13.a). It was developed with the constraints of fast and precise modelling grounds of any sizes. It is capable to process triangulation, to deform a meshed surface, to calculate contour lines and to bound plots for the identification of zones. This capacity to plan and to integrate an outline into the meshing of a ground is of major importance to refine the demarcation between plots for the affectation of objects, colours and textures. A polygonal description in real-time gives a first visualization of the terrain with shades (Fig.13.b). The terrain so to model will be then imported in IMAGIS.

Fig. 13 : Terrain modelling: points data (a) and rendering (b). Plot description IMAGIS can import data from GIS as ARCINFO© for example. Here, the data are coming from the Vosges mountains and were supplied by ONF1. The left map (Fig.14.a) represents the distribution of main species on the considered zone. The right map (Fig.14.b) represents a state of the plots of land after the passage of the storm in 1999. Every colour corresponds to a percentage of brought down trees.

Fig. 14 : Distribution of main forested essences (a) and effect of the storm in 1999 (b).

1 Special thanks with ONF and Christian PIEDALLU (ENGREF)

(a) (b)

(a) (b)

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 590

Fig. 15 : Plot visualisation in IMAGIS after GIS data importation.

The figure 15 represents the description of the plots after the import of the data in IMAGIS. Specific menus allow then to describe every plot: species, density of plantation... It is by means of these tools that are also associated to every plot trees generated by AMAP. Tree generation and 3D visualisation The simulation of 3D shapes of the trees is done by the module of the AMAP chain called Genesis. This is a procedural growth engine deriving from the AMAP research growth simulator and designed for plants and trees. It can be used to simulate a vegetable in 3D at a given age by taking into account the natural morphology of the plant. A big number of features allows to control the aesthetics and the simplification of the tree, to control its botanical behaviour during time, to prune it, to parameterise the seasons, the colours and the textures of leaves and trunks, to control the number of polygons exported according to the distance and to generate different geometries for the same vegetable. The Landmaker workshop can be considered as the "conductor" of the AMAP chain: it allows to gather terrains, buildings, infrastructures and vegetations in the same virtual scene (Fig.16). One of its particularities with regard to the other existing software is that there is no limitation on the number of polygons composing the scene to be treated. It is indeed capable of integrating, staging and showing virtual models of big landscapes. Specific tools allow the plantation of big zones of vegetation or objects in the form of adaptations, geometrical plantations, and wild copses. It is possible to associate different species to each zone with various densities of plantation.

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 591

Fig. 16 : Landmaker module: top view and 3D view (up left part) of the plots .

So, by using Genesis's capacity to generate trees and those of Landmaker to show the landscape, IMAGIS allows to represent in 3D the 2D data of a GIS (Fig.17-18).

Fig. 17 : Realistic views from a plot to different ages. In this application, we contented with showing a part of Vosges from GIS data. The next step consists in making a link between this 3D reconstruction and the data supplied by software taking into account forest models, such as WinEPIFN. It will be so possible to reach as well the architecture of the trees as their forested characteristics: biomass, quality of the wood, etc. THE SILVES PROJECT General presentation The resolution of the problems of management of forest resources requires to assemble complex spatio-temporal processes: integration of the knowledge in software of simulation, extended use of GIS, purchase the economic or ecological optimisation of the forestry. Within the framework of a collaboration inter CIRAD-INRA-ENGREF-INRIA institutes, we try to develop tools allowing to help in the management of plantings by using GIS, the models of forest production of the INRA, the models structure - function of CIRAD-AMAP and the knowledge in high performance computing and visualization of the ISA project (applications for Increased Reality in real time) in INRIA-Lorraine. In particular, modelling a 3D forest cover from GIS data requires numerous treatments: developing interactive visualization systems, as a tool of dialogue between the administrators and the users of the forest, calculating aerodynamic and transfer of energy within the forest cover, etc. Thus, within the framework of ISA project, the scientific project SILVES would result in an integrated software platform to represent, simulate and visualize forestry spaces in their current state and during their evolution in order to better control the management decisions. The geometrical modelling of forest covers could also be used by others models interested by the structure of forests: simulation of forest fires, simulation of the wind within the covers and resistance to breakage, evaluation of the wood quality according to the forestry constraints, etc. Research topics Scientific objectives of project SILVES can be divided into three principal research topics. 1. 3D visualization of GIS data Use of the software IMAGIS (visualization of GIS) and software AMAP (simulation of the growth of the plants) developed in CIRAD. In particular, it is necessary to develop the potentialities of IMAGIS (which was conceived originally for applications of landscape management) for a wider use in forest research topics. 2D images from GIS are the inputs of SILVES system. They are then processed for plant analysis (species, density, age, etc) and for the semi-automatic generation of 3D scenes with AMAP modelling and simulation techniques. 2D images from IMAGIS GIS, when combined with DEM (Digital Elevation Models) data, can also be used as background in bird-eye's views of 3D scenes.

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 592

2. Coupling structure-function models and forest production models The main idea is to develop an interactive relationship between structure-function models and forest production models in order to pull together advantages of each system: in one hand, the tree architecture is well described but the stand structure is not managed, on the other hand, forest production is estimated at broad scales but the tree description is very simplified. Coupling both approaches would then offer realistic simulations on a large area, with a good description at the stand and the tree level (including tree architecture, wood quality etc.). The goal is to reconstruct completely the plant in 3D, throughout its life and with precision knowing the general rules of development of the species and knowing only some variables relative to a realization at a given time of the plant development. This problem of inversion of models and quasi-numeric is similar to the estimation of parameters or to the optimisation process.

3. High performance Computing and visualization Simulation of forest covers and visualization in immersed room. The characteristics of this topic are as follows: • Development of a software platform allowing to simulate the development of a forest cover (by using existing

competences of AMAP) and to visualize it in immersed room. • Simulations in real time in order to allow an interaction between the user and the simulator (by using high

performance virtual reality hardware and most recently developed techniques in computer graphics, such as efficient shadow map and combination of point, line, and volume models).

• Study of the multi-scale representation of a forest cover so to optimise the performances. (by using level of detail and model simplification techniques).

• Development of an unified interface allowing to modify the characteristics of the simulation (forestry choices for example) with AMAP and IMAGIS supported as back-end.

CONCLUSION AND OUTLINES The integration of a management plan in a GIS and the ability of computer graphic systems to simulate plant mock-ups offers the possibility to predict the evolution of a landscape with time, and to project different scenarios. With the advent of computer graphics and virtual reality techniques, land-use managers have become interested in such tools which help represent rural landscapes. The development and the use of such computer tools show that the representation of geographic data provides by a GIS through virtual computer graphic images can provide a better understanding of the landscape by a non-specialist public. The resulting simulations lead to realistic images, which can become efficient aids to facilitate understanding, communication and decision making by the various stakeholders in a land management project. The researches led by the program AMAP and by the LERFoB, on the growth of plants, notably forested trees, supply methods and knowledge which it is then possible to assemble in quantitative tools of management. We aim on one hand to understand, to predict and to simulate the functioning and the production of the plant stands (forested, agricultural and agroforested) and on the other hand to simulate rural or urban landscapes to reconstitute the evolution of the environment, represent the consequences of organizations and analyse the impact of the strategies of land management. The SILVES (Computer simulations and visualisation software applied to sylviculture) project was defined to allow the software development of an integrated platform which can be used in a forested frame. REFERENCES Auclair D., Barczi J.F., Borne F., Etienne M.., 2000: Landscape visualisation software as a forest management decision support system. Proc. Int. Conf. Criteria and indicators for sustainable forest management at the forest management unit level, Nancy, France, 21-25/03/00. Barczi J.F., Borne F., Perrin L., 2000 : SIG et modélisation procédurale du paysage : l’aménagement du territoire dans ses trois dimensions. UrbAO , 1:40-43. Bionatics®, 2000: Bionatics : Born from the AMAP research. Web Site, http://www.bionatics.com. Blaise F., Barczi J.F., Jaeger M., Dinouard P., Reffye (de) P., 1998: Simulation of the growth of plants – Modeling of metamorphosis and spatial interactions in the architecture and development of plants. In : Cyberworlds, Kunii T.L., Luciani A. Eds..

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 593

Blaise F., Houllier F., Reffye (de) P., 1996: Simulation of tree architecture and growth in a forest stand: AMAPpara software. In Connection between silviculture and wood quality through modelling approaches and simulation softwares IUFRO WP S5.01.04 Workshop, Hook, Sweden. Ed. G Nepveu INRA, 46-55. CIRAD- AMIS , 2000 : CAPSIS : Simuler la croissance et la sylviculture. Web Site, http://www.cirad.fr/presentation/programmes/amap/logiciels/capsis.shtml. Courbaud B., Goreaud F., Dreyfus P., Bonnet F.-R., 2001: Evaluating thinning strategies using a Tree Distance Dependent Growth Model : some examples based on the CAPSIS software « Uneven-Aged Spruce Forests » module. Forest Ecology and Management 145, 15-28. Dauzat J., Eroy N.M., 1996: Simulating light regime and intercrop yields in coconut based farming systems. European Society for Agronomy, 7-11 July, Wageningen, The Netherlands, 16 pp.

Dhôte J.-F., 1995 : Définition de scénarios d’éclaircie pour le Hêtre et le Chêne. Revue Forestière Française XLVII, 106-110.

Dreyfus P., 1993 : Modelling Austrian black pine response to silvicultural practices in the South East of France. In : Burkhart H.E., Gregoire T.G., Smith J.L. (eds.) Proceedings of the IUFRO S4.01 Conference « Modelling Stand Response To Silvicultural Practices », sept 27-oct 1, 1993, Publication FWS-1-93, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA, 5-18. Gary C., Barczi J.F., Bertin N., Tchamitchian M., 1994 : Simulation interactive de la croissance d’une plante de tomate en fonction de son environnement et de sa conduite (modèle TOMGRO). In: Architecture des arbres fruitiers et forestiers, Les Colloques I.N.R.A. Edition , 1994, 333-343.

Gourlet-Fleury S., 1999: Individual-based spatially explicit modelling of forest stands in French Guiana. In : Y. Laumonier, B. King, C. Legg and K. Renolls (eds.) Proceedings of the international conference “Data Management and Modelling Using Remote Sensing and GIS for Tropical Forest Land Inventory”, october 26-29, 1998. Jakarta, Indonesia, 473-490. Gourlet-Fleury S., Houllier F., 2000 : Modelling diameter increment in a lowland evergreen rain forest in French Guiana. Forest Ecology and Management 131(1-3), 269-289. Houllier F., Leban J.M., Colin F., 1995. Linking growth modelling to timber quality assessment for Norway spruce. Forest Ecology and Management, 1995, 74, 91-102. IMAGIS®, 1996 : IMAGIS. Web Site, http://imagis.cirad.fr. Loustau D., Pluviaud F., Bosc A., Porté A., Berbigier P., Déqué M., Pérarnaud V., 2001: Sub-regional climate change impacts on the water balance, carbon babalance and primary productivity of maritime pine in south-west France. In: Models for the sustainable management of temperate plantation forests, EFI Proceedings N°41. J-M. Carnus, R. Dewar, Loustau D., Tomé M., Orazio C. (eds .), 45-58. Leban J.M., Daquitaine R., Saint-Andre L., 1996. Un outil d’évaluation de la qualité de la ressource en bois appliqué au Douglas : Le logiciel Win-EPIFN. Forêt Entreprise,109, 11-15. Leersnijder P., 1997: The Pinogram program : PINe Growth Area Model. Web Site, http://ourworld.compuserve.com/homepages/Peter_leersnijder/pinogram.htm. Lintermann B., Deussen O., 2000: Greenworks, organic-software. Web Site, http://www.greenworks.de/greenWeb2/Templates/start.html. Mäkela A., Mäkinen H., Vanninen P., 1999: Quality of scots pine stems under different thinning treatments. An analysis based on resource capture and allocation in idividual trees. In : Gérard Nepveu editor, Proceedings of the Third Workshop “ Connection between silviculture and wood quality through modelling approaches and simulation software ” La Londe-les-Maures, France, September 5-12, 1999, INRA-Nancy, 571-578.

Fourth Workshop IUFRO S5.01.04, Harrison Hot Springs, British Columbia, Canada. Proceedings edited by G. Nepveu

__________________________________________________________________________________________________________

������������������ 594

Maujé, 1975 : Modèle de croissance et de production des peuplements modernes de Pin Maritime. Rapport annuel de l’AFOCEL, 227-249.

Meredieu C., 1998 : Croissance et branchaison du Pin laricio (Pinus nigra Arn. ssp. laricio (Poir.) Maire) : Elaboration et évaluation d’un système de modèles pour la prévision de caractéristiques des arbres et du bois. PhD Thesis, University Claude Bernard Lyon I, 238pp. + app. Prusinkiewicz P., Hammel M., Mech R., 1994: Visual Models of Morphogenesis : A Guide Tour. Web Site, http://www.cpsc.ucalgary.ca/Redirect/bmv/vmm-deluxe/TitlePage.html.

Reffye (de) P., Houllier F., 1997: Modelling plant growth and architecture: Some recent advances and Applications to agronomy and forestry. Current Sciences, vol. 73. Saint-André L., Laclau J-P., Bouillet J-P., Deleporte P., Miabala A., Ognouabi N., Baillères H., Nouvellon Y., 2002: Integrative modelling approach to assess the sustainability of the eucalyptus plantations in Congo. Fourth workshop, IUFRO Working Party S5.01.04 « Connection between Forest Resources and Wood Quality: Modelling Approaches and Simulation Software». Harrison Hot Springs Resort, British Columbia, Canada, September 8-15, 10pp. A mon père, René BLAISE (4 avril 1931 – 12 août 2002)