MAppleT: simulation of apple tree development using mixed stochastic and biomechanical models

Area of Science:

• * Computational botany
• * Plant modeling
• * Horticultural science

Background:

• * Traditional tree architectural databases are time-consuming and struggle to capture dynamic changes in topology and geometry.
• * Simulating tree development requires integrating models of both plant structure (topology) and physical form (geometry).

Purpose of the Study:

• * To develop a unified simulation approach for modeling apple tree architecture by integrating topology and geometry.
• * To create a dynamic 3D model that captures the emergence of tree form from process interactions.
• * To assess the model's accuracy by comparing simulated and digitized tree characteristics.

Main Methods:

• * Utilized L-systems to integrate topological and geometrical modeling.
• * Employed a mixed approach with stochastic models for plant topology and mechanistic models for geometry.
• * Modeled growth units (GUs) and branching using a hierarchical hidden Markov model.
• * Incorporated a biomechanical model to simulate stem form, considering growth dynamics and gravity.

Main Results:

• * Developed MAppleT, a simulation tool generating 3D mock-ups of apple tree progression over time.
• * Performed sensitivity analysis and compared simulated versus digitized tree descriptors (GU count, shoot/axis geometry).
• * Demonstrated the model's ability to represent key architectural features and growth dynamics.

Conclusions:

• * MAppleT successfully integrates topology and geometry for simulating apple tree development.
• * The model captures the influence of gravity on tree architecture.
• * Despite limitations, MAppleT is a valuable tool for understanding and predicting apple tree growth patterns.