Institute of Forest Biometry and Informatics
Faculty of Forest Sciences and Forest Ecology
at the
University of Göttingen
PLANT MODELLING GROUP
PROJECT:
Development of statistical and geometric methods for the
characterization of the architecture of plants and
forest stands and for comparing models
Short description of project
For the purpose of validation and comparison of 3-dimensional tree and
stand models, new statistical and geometric-combinatorial analysis tools
shall be developed and implemented in a scientific software, together
with established methods. Points of departure are geostatistical
methods, Markoff chains, multivariate distributions and
geometric-combinatorial characteristics (e.g. neighbourhood relations
in tree stands). Aims with practical relevance are: (a) the precise
characterization of complex 3-dimensional plant ensembles in spatially
heterogeneous (mixed-species) forest stands, (b) upscaling of models,
i.e. the extraction of relevant information when the level of detail
is coarsened, (c) taking a property depending on tree architecture
into account in forest stand models - an example of such a property is
wood quality. The dynamic character of the simulated plant structures
and the fundamental role of physical and physiological mechanisms
in tree growth and competition shall be taken into account.
Aims
(a) A "toolchest" of methods devoted to the description and comparison
of single-tree and stand structure shall be developed (theoretical
description, implementation and test). New as well as established
methods shall be used. The methods for comparing datasets shall be
designed for comparison of simulation results with reality (model
validation) as well as for comparison of model results with each
other (taking models from other research groups into account).
(b) Interrelations between several methods of description shall be
subject of a mathematical analysis (analytically, or by model
experiments). E.g., the relationships between the following
approaches are until now not well clarified in mathematical
terms:
- causally-based models of competition (e.g. using the notion of
shadowing biomass),
- marked point processes,
- geometric-combinatorial characteristics (configurations of the
neighbourhood of trees),
- univariate and multivariate distributions of different indices
of distance and density,
and for the single tree also:
- physiologically-based models of growth and allocation,
- the stochastic (branching-oriented) approach of the AMAP model,
- grid-based indices (e.g. box counting dimension).
(Parts of these mathematical investigations will be carried out
by
Winfried Kurth.)
(c) In the form of examples (two or three cases), a "reduction of
models" shall be carried out: From detailed models (LIGNUM, AMAP,
possibly MADEIRA) which are more or less physiologically-based,
simplified, rule-based models shall be derived which are less
demanding in terms of calculation time and can therefore be applied
for larger tree stands. The challenge is to preserve the essential
characteristics of growth dynamics, competition and development of
structure during this simplification process.
Steps
- Preliminary and accompanying technical work at the software tools
for tree and stand representation, modelling and analysis,
including the implementation of a data interface between the
simulation tools LIGNUM (from
METLA) and GROGRA,
- design and implementation of analysis tools,
- investigation of the performance of the implemented tools and
of their interrelations,
- possible application to the assessment of wood quality in structured
tree stands.
A research stay of
Helge Dzierzon at the Finnish Forest Research Centre has already
served to prepare the data exchange with the Finnish group and to
get acquainted with problems emerging from the different treatment
of tree stands and individual trees in models. A software tool for
treating spatially anisotopic competition effects
as they influence the growth of a single tree in the LIGNUM model
was developed.
Cooperation partners:
Dr. H. Spellmann and Dipl.-Forstw. M. Guericke,
Department of Growth and Yield at the Forest Research Station of
Lower Saxony, Göttingen,
Dr. Risto Sievänen and Dr. Jari Perttunen
at the Finnish
Forest Research Centre (METLA), Finland,
Prof. Dr. Hans Pretzsch,
Chair for
Forest Growth Modelling, Technical University of Munich
Start of project: March 15, 2000.
End of project: February 28, 2003.
Funding: DFG project.
Intermediate report (in German)
Final report (in German)
This project was carried out by
Helge Dzierzon.
Other projects of the research group
Back to Plant Modelling Group Homepage
Last modifications: August 13, 2003