Institute of Forest Biometry and Informatics
Faculty of Forest Sciences and Forest Ecology
at the
University of Göttingen
PLANT MODELLING GROUP
PROJECT:
SPATIALLY EXPLICIT, PHYSIOLOGY-BASED MODELS OF
WOODY PLANTS
1. Development of a formal language for spatially explicit modelling
and simulation of the allocation of carbon and nitrogen and of growth
in woody plants
The central aim of the project is the development of a general formal
language designed to describe C and N allocation and growth in woody
plants, particularly in forest trees. A software
which "understands"
this language and which generates 3D simulation models of plants
will be a natural by-product.
Lindenmayer-systems (L-systems) will serve as a basis. These are
string-rewriting systems; each L-system specifies a set of replacement
rules describing the evolution of strings which encode 3D branching
structures. "Stochastic sensitive growth grammars" are an extended
version of L-systems, they are able to represent features of plants like
reaction to overshadowing or density, and also already some simple forms
of carbon allocation modelling. A software developed at our institute,
called
GROGRA
(Growth Grammar interpreter), can read these grammars
from text files and generate time series of morphological (3D) structures
representing the plants or the forest stand. GROGRA contains also
several output filters for standard graphical file formats (PostScript,
HTML, PovRay, DXF, AMAP) and some analysis tools to explore the
virtual plants and to compare them with empirical data from
measured plants. More information on this preliminary work and
descriptions of the growth grammar language and of the software
can be found in
the
publications of the research group.
Of course it is possible to construct simulation models of plant
architecture, growth and carbon metabolism directly in some
standard programming language, without using a formal description tool
like "growth grammars", and this was in fact already done by many
research groups in ecophysiology, agricultural or forestry-related
projects. However, using grammars as a means for specification has
some advantages:
- The description in terms of grammars is precise and relatively short,
- it enhances transparency and reproducibility,
- a formal specification allows for an abstraction from unimportant
details, thus comparisons between different models give much more
evidence when the models are given in formal terms,
- models can much easier be combined with each other in this case,
- general theoretical tools (e.g. theorems from formal language
theory) become applicable.
The "stochastic sensitive growth
grammars" have still some weaknesses,
particularly when the physiological basis of plant growth is to be
taken into account. Sensitivity of processes to environmental influences,
which is represented in the current
grammars by specialized "sensitive
functions", should be modelled in a much more general and transparent
way. Likewise, the formal tools to describe allocation and plant-internal
control mechanisms are not yet optimized.
The exploration of the physiological background and of control
mechanisms governing the allocation of carbon and nutrients and the
onset of growth in plants is one of the central topics of plant
physiology, and there exists already a huge number of diverse modelling
approaches. The formalism which shall be developed here should be
able to represent several of these models (to allow for precise
comparisons). Model experiments and sensitivity tests can then
perhaps help to clarify some of the questions concerning mechanisms
of allocation discussed in the literature.
Methodologically, a combination between the rule-based
and the object-oriented programming paradigm will be necessary
for the specification language.
2. Simulation of the reaction of plants to competition and
pruning as a result of the interaction of plants and plant organs
with each other
The object-based approach, which considers the plant as a complex
3D structure composed of relatively simple units (leaves, internodes,
growth units, root segments etc.), has the advantage that changements
in architecture which are caused by competitors or by local damage
can appear in the model as emergent phenomena, resulting from the
interaction of the units, and need not be introduced artificially.
Manipulations in forestry as well as the pressure exerted by
competitors are cases where spatially explicit models can integrate
the knowledge which is currently available at the scale of single
plant organs, and where the consequences for the whole plant or
for tree stands can be checked against reality. Modelling thus becomes
a means to explore the consequences which assumptions introduced at a low
level can have at a high level (several interacting plants, a forest
stand).
3. Taking into account some fundamental results concerning genetic
control of plant morphogenesis
Ecophysiological models of plant growth concentrate normally on
the environmental factors which have influence on allocation and growth,
and they emphasize the opportunistic reactions of plants to changements
in these factors. However, the genetically
fixed "blueprints" of the
plant organism are equally important. Endogenous control of the
activity of meristems during ontogenesis determines the shape of the
resulting plant to a considerable extent. This is the reason why
we can distinguish e.g. a birch tree from an oak tree by their habit.
Basically, L-systems are designed to represent a development of plant
form which is genetically determined. However, until now they do not take
some recent results of plant genetic into account. During ontogenesis,
some genes are activated in a spatially and temporally limited manner.
Here, transcription factors activating or repressing the expression
of specific target genes play an important role. The ability to
represent regulating networks of such factors would enhance the
usefulness and power of growth grammars in the field of plant morphology
considerably. It is therefore planned to cooperate with other research
groups to extend the formalism of growth grammars in a way which
opens the possibility to take such networks into account.
4. Development of statistical and geometric methods to characterize
the spatial structure of plants and plant stands and to compare
results of models
For a validation of the results of new models, a methodologically
reliable comparison of simulation results with empirical data is a
prerequisite. Furthermore, the comparison between the results of several
models will be of growing importance in the future. Methods of evaluation
and comparison which are adapted to the spatially explicit models
developed here are subject of a
related project.
Start of project: January 1, 2000.
End of project: September 30, 2001.
The project was originally planned for 5 years. Premature interruption
occurred because of W. Kurth's call to a tenureship at the Technical
University of Cottbus. However, important parts of the project will
be continued further, at Göttingen and Cottbus.
Funding: Heisenberg scholarship of the
DFG.
Final report (in German)
This project was led by
Winfried Kurth.
Other projects of the research group
Back to Plant Modelling Group Homepage
Last modifications: August 13, 2003