OBJECT MODELING AND SCALING OF LANDSCAPE PROCESSES AND CONSERVATION EFFECTS IN AGRICULTURAL SYSTEMS
Location: Agricultural Systems Research Unit
Title: CANON in D(esign): Composing Scales of Plant Canopies From Phytomers to Whole-Plants Using the Composite Design Pattern
Submitted to: Wageningen Journal of Life Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: August 27, 2009
Publication Date: December 9, 2009
Citation: Mcmaster, G.S., Hargreaves, J.G. 2009. CANON in D(esign): Composing Scales of Plant Canopies From Phytomers to Whole-Plants Using the Composite Design Pattern. Wageningen Journal of Life Sciences. 57:39-51. Available: http://dx.doi.org/10.1016/j.njas2009.07.008.
Interpretive Summary: Since Grey first presented the phytomer concept in 1879, it has been useful for understanding plant development and architecture. Most people have viewed the phytomer as a vegetative unit consisting of a leaf, node, internode, and axillary bud, and some scientists have extended this definition to include the root nodal bud and the inflorescence structure. For many reasons, crop simulation models do not fully incorporate the phytomer concept in the approach. The objectives of this paper are to provide detail on the translation of the botanical abstraction of a phytomer into OO design and to present a proof-of-concept prototype OO plant model, CANON based on winter wheat as a case study. We named the model CANON because the interplay of repeating phytomers is analogous to the repeating melodies of a musical composition called a canon.
This paper describes how the use of a Composite Pattern in an object-oriented (OO) design facilitates implementation of different scales from the phytomer to a mixture of single and aggregated phytomers for different plant components into designing simulation models. In our proof-of-concept prototype of CANON, we demonstrate that our approach allows specification of sub-models of different levels of detail which can be selected at implementation or run-time. For instance, we simulated the development of the entire wheat canopy by the dynamic interplay of phytomers from planting to maturity, including the head. We also substituted alternative code for aggregating the phytomers into a single head, thereby demonstrating the ability to interchange code at different levels of scale and complexity. The phytomer concept of building plants meshes naturally with the Composite Design Pattern which, by providing a consistent interface between components (phytomers or aggregates), simplifies the coding and reduces potential for errors in developing and maintaining simulation models. The design provides open-ended recursion of phytomers, which facilitates easy implementation of different determinate and indeterminate plant species. This approach also provides new opportunities in modular modelling approaches such as the Object Modeling System.
The phytomer concept has been useful for understanding plant development and architecture. Normally the phytomer has been considered as a vegetative unit consisting of a leaf, node, internode, axillary bud, and occasionally nodal roots, although the definition can be extended to the inflorescence structure. These units are repeated within and among shoots. Whether because of lack of knowledge or specific modelling objectives, crop models may not fully incorporate phytomer concepts, rather some phytomers may be aggregated into a single component such as a grain or inflorescence component. The continuing development and maturing of object-oriented (OO) design and associated programming languages is providing opportunities for better representing phytomer concepts in crop simulation models. For instance, use of the Composite Design Pattern in an OO plant design facilitates implementation of different scales from the phytomer to a mixture of single and aggregated phytomers for different plant components. This paper uses winter wheat (Triticum aestivum L.) to illustrate 1) how plants build their canopies by the appearance, growth, and abortion/senescence of phytomer units, and 2) presents a prototype translating this botanical abstraction into an OO design of a plant model, CANON, named because the interplay of repeating phytomers is analogous to the repeating melodies of a musical composition called a canon.
Development of CANON involved the implementation of the phytomer concept (both vegetative and reproductive phytomers) into an OO design. All phytomers required a minimum of an internode and axillary bud components. Vegetative phytomers added leaf and nodal root components, while reproductive phytomers added floret/kernel components, although depending on the situation, they may not be present for a specific phytomer. Simplified developmental and growth rules primarily derived from the SHOOTGRO and APSIM-Wheat (v53) models were used for proof of concept of CANON. The structural composite design pattern provided the ability to simulate scales from individual reproductive phytomers to an aggregated spike on a shoot (or aggregated into entire plant). CANON provides a template for an OO design to simulate plant canopy development and growth from the subphytomer level to whole-plant aggregated level, thereby allowing flexibility in meeting different model objectives and available knowledge.