Institute of Forest Biometry and Informatics

Faculty of Forest Sciences and Forest Ecology
at the
University of Göttingen

Short description of project

Existing process-oriented and structural models of trees shall be linked together in such a way that process-related information and structural information can actualize each other. To achieve this aim, certain approaches from computer science shall be applied: rule-based systems, object-oriented simulation, communicating processes. A mediation between different modelling concepts and between different scales shall be realized. The prototypical application will be the modelling of radiation interception, transpiration and architectural development of trees. Three basic models will be coupled together: A structure-oriented growth model of tree crowns, a model of microclimate inside the crown, and a model of water flow in the tree. The technical realization will be carried out in a generic way, i.e. each of the used basic models can be substituted by another one which simulates the same phenomenon (although possibly in another way, or at a differing level of detail). This linkage of models shall contribute to a better understanding of the influence of spatial heterogeneity in forest stands on the hydraulic regime of the trees and on their resistance to drought stress.


Aims

(a) Models of microclimate (solar radiation, temperature, water vapor in the air, wind speed) shall be connected with models of tree architecture (branching structure of the crown) and water flow in the tree, with information flows from the structural model to the microclimatic model, and also from the environment (microclimate) and from the structural model (architecture) to the water flow simulation. This will enable the calculation of a spatially heterogeneous transpiration flux density and its use as an input for the water flow model. In model experiments it will be assessed which consequences this heterogeneity in the crown will have for the dynamics of water flow and for the risk of drought-stress induced damage (cf. project on water flow simulation). Furthermore, such a triple interaction of models (microclimate / architecture / flow dynamics) is a prerequisite if stomatal control shall be included in structure-oriented tree models.

(b) It is also planned to realize a "backward" information flow: From microclimate and water relations to the growth model which generates the tree architectures. This will allow some tests of hypotheses by model experiments, concerning the question how the processes of PAR interception and transpiration (resp., their disturbance by shadow or by drought) influence the control of shoot growth in the tree crown.
It is not planned to implement a full physiology-based carbon fixation and allocation model for this purpose. Rather, heuristic sensitive growth rules will reflect the observed responses of trees on changements of radiation regime or water supply. A refinement of these rules to obtain a higher physiological realism can be approached later (cf. project on physiology-based models of woody plants). In any case, the model experiments can ascertain the inner consistency of the used model components and of their interaction, and they will give information about sensitivity against changes of parameters and / or model structure.

(c) Genericness and transparency of modelling tools (process-oriented and structure-based) shall be improved. We have therefore begun to reimplement some model components in an object-oriented language (C++) and to use standard class and template libraries which can be shared by several groups of researchers. It is also planned to develop a generic software tool for discretization of branching structures and environment according to different criteria. To foster a broad dissemination of this tool and of the model-connecting software, graphical user interfaces shall be implemented for these software systems, using standardized libraries and tools which are easily accessible for interested modellers in other institutions.


Existing models which are used

Microclimate in tree stands: The two models MIR and MUSC were partially reimplemented in this project and will be utilized for the calculation of spatially heterogeneous transpiration rate. These models were developed by Dr. Jean Dauzat at CIRAD-amis, Montpellier; see the Project "AMAPlux". MIR calculates the incoming radiation and MUSC takes multiple scattering into account. Furthermore, the energy balance of the leaf as well as stomatal conductivity are simulated in separate modules also developed by J. Dauzat. There will also be the option to substitute MIR/MUSC by a different radiation model developed by Y. Knyazikhin and O. Panfyorov at the Institute of Bioclimatology, University of Göttingen.

Tree architecture and growth: Alternatively, the models AMAPsimpar (cf. AMAP projects at the CIRAD, Montpellier) and GROGRA are used for modelling the spatial architecture of the branching system of the tree crown and its temporal development, including the growth of the tree.

Water flow in the tree: Alternatively, the models HYDRO, developed by J. Dauzat at CIRAD-amis in connection with the microclimate models (see above), cf. Project "AMAPlux", and HYDRA (with higher temporal resolution of flow dynamics), developed by Th. Früh in our institute and currently being subject of the project on numerical simulation of the hydraulic system of trees, will be used.

Since all these simulation tools are advanced systems with a rather complex functionality and operating on heterogeneous 3D-structures, the design and implementation of data exchange interfaces between them is nontrivial. It requires certain steps of standardization, but we avoid the complete reimplementation of the models. Some of the models will get a clearer modular structure, others will only be extended by normed data interfaces for communication.


Working programme

Tasks already done:

• During a research stay at the CIRAD (Montpellier) in the team of Dr. Dauzat, a concept for modularization of microclimatic and tree-hydraulic modelling has been developed, and essential parts of the models MIR, MUSC and HYDRO have been reimplemented. This has turned out to be necessary in order to open the possibility to substitute parts of the compound model system by other, alternative models (as it is planned for HYDRO / HYDRA). In the old version, the components of the microclimatic and tree-physiological models have been interwoven in a way that has excluded such a modular treatment of program parts. In order to enhance modularity, transparency and genericness of the code, an object-oriented design was chosen and implemented in C++.
• Parallel to this reimplementation, a new software module, called NEXUS, was built which will control the communication between the different microclimatic, physiological and structural modules. The NEXUS module is currently being finalized for platform-independent use (i.e. under Unix as well as under MS-Windows). It contains also algorithms for manipulation of structural tree data which have been developed by D. Lanwert (Göttingen) during a preceding, EU-funded project at the CIRAD (Montpellier).
• A concept for a "backward" information flow from the numerical simulation model HYDRA to the growth engine of AMAPsimpar (based on the reference axis principle) has likewise been developed in the frame of that preliminary project in a joint work by J.-F. Barczi (CIRAD) and D. Lanwert (University of Göttingen).
• Sample trees and branches of durmast oak (Quercus petraea) have been measured and represented in digital form in order to have test data for the model. This data acquisition work has been done together with the project on tree hydraulic function.

• Finalization and tests of the new, resp. reimplemented model components MIR, MUSC, HYDRO, NEXUS.
• Inclusion of the hydraulic water flow model HYDRA (with spatially heterogeneous transpiration flow) into the new model system as an alternative to HYDRO, and tests of the combined model.
• Systematic model experiments with the new system to evaluate the performance and to compare the results with previous model runs (based on the old software) and with empirical data.
• Improvement of the data interface to AMAP, based on the new standard for architectural tree data (MTGs = multiscaled tree graphs).
• Implementation of standardized information pathways from microclimatic and physiological models to the growth engines of AMAPsimpar (already prepared) and of GROGRA.
• Development of a generic and user-friendly software tool (with graphical interface) for topological and geometrical discretization schemes and for their application to virtual trees (GroDisc programme), and its linkage to NEXUS.

Final report (in German)

This project was carried out by Gustavo Alejandro Anzola Jürgenson.

Other projects of the research group

Back to Plant Modelling Group Homepage