The Universal Individual-Based Model (UIBM)
Final Version
A joint project of the
Institute for Plant Ecology,
Department of Biology, Chemistry and Earth
Sciences
Justus-Liebig-University
and the
Department of Mathematics, Natural Sciences and
Computer Science,
University of Applied Sciences
Model Description
1. Overview
We are in a phase of a dramatic, unprecedented
loss in biodiversity on earth (1,2). The direct drivers
of biodiversity loss are land cover change (1), habitat change (1), invasive
alien species (1), pollution by excessive usage of synthetic nitrogen
fertilizer (1,3) and by nitrogen deposition from the
atmosphere (4), global climate change (1,3,4), and rising atmospheric carbon
dioxide levels (4). In principle, phytodiversity
loss, i.e. the decline in plant species richness, is a crucial component of the
man-made biodiversity loss.
1.1 Model Purpose
UIBM will be an agent-based discrete-event
simulation model of the dynamics within multi-species plant communities. The
aim of the Universal Individual-Based Model is to provide a modeling tool for
understanding and prediction of phytodiversity loss
in
1.2 Current Model Structure, State Variables
and Scales
We have begun to construct the architecture of a template
species, i.e. false oatgrass (Arrhenatherum
elatius), mainly from online species traitbases of the Northwest European Flora. Due to its
development along traitbases, in principle, the model
is extensible to the whole set of herbaceous species contained in those
databases.
We have
used Java together with the Multiagent Simulation
Toolkit MASON (5) for the model development.
The main structure of the model is as follows:
1. Composite
Pattern (according to 24) for the plant components:
abstract superclass
PlantComponent
(instance
variables: location, direction, rollPitchYaw)
â â
abstract superclass:
PlantStructComponent concrete
subclass Meristem
(dry mass,
carbon/nitrogen concentration, content) (meristem type)
â
concrete subclasses
Vegetation, PlantIndividual,
Plant (newly attained carbon/nitrogen per day,
since last organ creation step),
BaseCoarseRoots (diameter/depth of rooting system),
BaseFineRoots (diameter/depth of rooting system),
CoarseRoots, FineRoots
(tissue density, nitrogen concentration, specific root length), ClonalGrowthOrgan, Internode
(length/diameter, volume fraction of living tissue in hollow
organs, density of living tissue, tissue
density)
Leaf (length/width, dry matter content, tissue
density, thickness, leaf mass per area, area), Flower (number of pollen, number
of seeds)
Pollen (number of pollen),
Seed
Within UIBM plants are modeled as ramets composed of internodes, leaves, coarse/fine root
systems, and flowers with pollen/seeds. Plants are connected by clonal growth organs and the so-linked plants form a clonal plant individual. All these concrete structural
plant components possess dry mass, carbon/nitrogen content as state variables.
The various plant organs do have additional state variables that are listed in
the simple class diagram shown above. The meristems
of the various types, representing points of leaf/internode/flower
growth on the plant, are modeled mass-less, so the Meristem
class is a direct subclass of PlantComponent.
2. Data
Transfer Objects transporting species-specific information from the
parameter database (a Properties-like text file) to the
plants/internodes/roots: class SpeciesInfo (general
information related to the species), class AllocationInfo
(state-dependent information related to carbon allocation among foliage, stem, clonal growth organ, coarse roots, fine roots), class BranchingInfo (state-dependent information related to meristem numbers per phytomer and
meristem angles), class RootingInfo
(information related to the rooting system)
The parameters are mostly derived from traitbases (6-11), but also make usage of trait information
for a range of species contained in scientific publications (12,16,19).
3. State
Pattern representing behavior that varies with the ontogenetic stage of the
plant: interface StateInterface, abstract
The implemented methods compute carbon/nitrogen
allocation to organ types and subsequently to individual organs, calculate the
attributes of the new organs, and decide whether sufficient resources are
available to create new organs in a state-specific manner. In general,
accumulative and creational states alternate over time. The initial state of a
plant is set to an accumulative state (AccumulateToCotylState)
that leads to the formation of cotyledons, i.e. the first leaves of the
seedling (CreateCotylState). Plants are entering a
flowering state, when the month of flowering start given in the databases is
reached and the dry mass at that time is larger than the minimum mature plant
dry mass.
4. Data
Transfer Objects mediating between plants and their organs to be created:
class AttributesNewLeaf, class AttributesNewInternode,….
The attributes of the organs to be created are
dependent on newly attained carbon/nitrogen after applying a linear carbon
allocation scheme and reducing maximum organ nitrogen concentrations (obtained
from 16, 17) until the N demand matches N supply to the plant. The various
attributes for each plant organ are found applying so-called “universal”
scaling laws, i.e. power law allometric relationships
(14,15, using 16,17 and helper functions of 18) to interpolate between
minimum/maximum attributes. In principle, it is assumed that minimum organ
sizes are realized with minimum organ N concentration. The Attributes objects
transfer the set of instance variables to the respective organs.
In general, UIBM shall simulate the dynamics of
a multi-species plant community in a 4 x 4 x 3 m micro-world, a virtual
vegetation survey plot, so to say. The time scale currently used is 1 day. The
temporal extent is considered to be in the order of one or two decades.
1.3 Process Overview and Scheduling: Present
and Future
Currently, we are testing the accumulation and
creation routines using a universal scaling law (13) as growth function. The
function is parameterized with seedling growth rate and maximum/minimum plant
dry mass specific to the template species. The nitrogen supply is set to its
minimum/maximum. At present, simulation time proceeds in steps of 1 day. So
far, the processes taken into account are:
1.
Aboveground/belowground carbon allocation (linearly interpolated between
the maximum/minimum ratio of newly attained carbon to newly attained nitrogen)
2.
Carbon partitioning between foliage and stem (linearly interpolated
between maximum/minimum plant new C/new N ratio)
3.
Carbon partitioning between internode and clonal growth organ (assumed to be constant)
4.
Matching organ N concentrations to the plant N-supply and adjusting
organ attributes accordingly
Besides that, a group of us is working on the
2d visualization of the scene and on a 3d-visualization of the plant
architecture with computer-designed leaves and internodes, clonal
growth organs and rooting systems displayed as cylinders. The 3d-visualization
will make extended usage of the Java3d capabilities available in MASON.
Merely a tentative outline of processes to be
implemented in the future shall be given here:
- Light
Interception submodel
A light interception model is an
important prerequisite for cutting the dependence on the growth function and
making the leaf “structure” functional. The code we have already obtained from
M. Roehrig might serve as a suitable source for this
task (20). Since Roehrig’s light interception model
calculates he light regime within the canopy on a several-hour basis, a
corresponding reduction of the simulation time step is necessary. Functions for
day length and direct/diffuse potion of the sunlight have to be implemented.
2. Gas exchange submodel
A prototype for the leaf part of
this submodel including the Farquhar
& von Caemmerer photosynthesis description (21)
and transpiration has already been written. Since light, temperature and relative
humidity of the air are the driving forces to his submodel,
they have to be included. The computation of respiratory carbon losses imposed
by the various organs will be computed from their actual N concentrations, the
local temperature and some portion of the remaining carbon input. In order to
integrate these processes all organs have to implement a gas-exchange
interface. At the end of each day the results of the organ-specific carbon
budgets have to be pooled at the plant level. Hence, updating of the newly
attained carbon must pursue in the scheduling order.
3. Water-/Nitrogen-uptake submodel
This submodel
shall enable the rooting system to take up substances within the covered soil
volume thereby making root structure likewise functional. Competition among
rooting systems will be solved by a supply/demand routine. The BaseFineRoot class has to implement a corresponding
interface. Consequences to the scheduling are similar to those of the
gas-exchange process (see above). Precipitation has to be added as an
environmental factor and water potential has to be added to the state variables
of an organ.
4. Organ lifespan submodel
As in (25) the lifespan of the
various organs will be made dependent on their N concentration and the density
of the living tissue. When an organ is supposed to die it will likely make part
of its nitrogen content available to the newly attained nitrogen of the plant
before carbon allocation is scheduled.
Comparison with other modeling approaches
The completed model will differ from most
individual-based plant growth models (e.g. the JABOWA/FORET/SORTIE model
family, see 22) by integrating competition for all above-/belowground
resources. UIBM will be similar to the LEGOMODEL (23) with respect to its
foundation in database-derived parameters. Nonetheless, UIBM will be much
closer in appearance to the functional-structural type of plant growth models.
Major differences to similar approaches will lie in the application of scaling
laws to derive organ attributes and the model being developed with a standard
Agent-Based Modeling toolkit.
References:
1.
Millennium Ecosystem Assessment (2005): Ecosystems and Human Well-being:
Biodiversity Synthesis. World Resources Institute,
2.
Nature Insight (2000): Biodiversity Nature Vol. 405, No. 6783
3.
Sala O.E. et al. (2000): Review: Global Biodiversity
Scenarios for the Year 2100. Science Vol. 287. no. 5459, pp. 1770 - 1774
4.
Thomas C.D. et al. (2004): Extinction risk from climate change. Nature 427, pp.
145-148
5. Multiagent Simulation Toolkit MASON: http://cs.gmu.edu/~eclab/projects/mason/
6. Kleyer M. et al. (in prep.): The LEDA Traitbase: A database of plant life-history traits of North
West Europe. (online at: http://www.leda-traitbase.org/LEDAportal/)
7. Klimešová J. & Klimeš
L. (2006): CLO-PLA - a database of clonal growth in
plants (online at: http://www.butbn.cas.cz/clopla/
)
8. Flynn, S., Turner, R.M., and Stuppy, W.H.
2006. Seed Information Database (release 7.0, October 2006) http://www.kew.org/data/sid
9. Buchner R. & Weber M.
(2000 onwards). PalDat - a palynological database: Descriptions, illustrations,
identification, and information retrieval. http://www.paldat.org/
10. Fitter A.H. & Peat H.J. (1994): The Ecological Flora
Database. Journal of Ecology 82, pp. 415-425 (online at: http://www.york.ac.uk/res/ecoflora)
11. Klotz S. et al. (2002): BIOLFLOR – Eine Datenbank
mit biologisch-ökologischen Merkmalen zur Flora von Deutschland. Bundesamt für Naturschutz
12. Roulston T.H. et al. (2000): What governs protein content
of pollen: pollinator preferences, pollen pistil interactions, or phylogeny?
Ecol. Monogr. 70, pp. 617–643
13. West G.B.,
Brown J.H. Enquist B.J. (2001): A general model for
ontogenetic Growth. Nature Vol. 413, pp. 628-631
14. Wright I.J. et
al. (2004): The worldwide leaf economics spectrum. Nature Vol 428, pp. 821-827
15. Niklas K.J. (1995): Plant Height and the Properties of Some
Herbaceous Stems. Annals of Botany
75: pp. 133-142
16. Thompson K. et
al. (1997): A comparative study of leaf nutrient concentrations in a regional
herbaceous flora. New Phytologist, Vol. 136,
No. 4 pp. 679-689
17. Whtehead D.C. (2000): Nutrient Elements in Grassland. CABI
Publishing
18. Shipley B.
& Vu T. (2002): Dry matter content as a measure of dry matter
concentration in plants and their parts. New Phytologist Vol. 153 Iss. 2, pp. 359
19. de Jong M.
(2003):Reaktionen von drei Süßgrasarten mit unterschiedlichen Nährstoffansprüchen
auf erhöhte NH3-Konzentrationen und NH4+-Gaben in Rein- und in Mischkultur. Ph. D. Thesis,
Justus-Liebig-University
20. Roehrig M. et al. (1999): A
Three-Dimensional Approach to Modeling Light
Interception in Heterogeneous Canopies. Agronomy Journal 91: pp. 1024-1032
21. von Caemmerer S. (2000): Techniques in
Plant Sciences Vol. 2: Biochemical Models. CSIRO Publishing
22. Deutschman et al. (1997): Scaling from Trees to Forests: Analysis of a Complex Simulation Model.
Science Online (http://www.sciencemag.org/feature/data/deutschman/index.htm
)
23. Lehsten V. (2005): Functional analysis and modelling of
vegetation. Ph.D. thesis, Carl von Ossietzky Universität
24. Gamme E. (1995): Design Patterns. Elements
of Reusable Object-Oriented Software. Addison Wesley Longman
25. Moorecroft P.R. et al. (2001): A Method for Scaling Vegetation Dynamics: The Ecosystem Demography Model
(ED). Ecological Monographs,
Vol. 71, No. 4, pp. 557-585
CREDITS AND
REFERENCES
----------------------------------------
To refer to this
model in academic publications, please use:
Uwe Grueters, Roland Dahlem, Jochen Senkbeil, Markus Woetzel (2007):
The
Universal Individual-Based Model (UIBM).
http://sourceforge.net/projects/uibm-de
In other pudeblications,
please use: © Copyright. April 20, 2007.
Uwe Grueters, Roland Dahlem, Jochen Senkbeil, Markus Woetzel.
Some rights
reserved.
See http://sourceforge.net/projects/uibm-de
for terms of use.