From greenhouse climate to individual leaves: an organ-resolved model of lettuce growth

arXiv cs.GR details an organ-resolved lettuce growth model that moves beyond whole-plant averages and simulates each leaf as its own physiological unit. The framework assigns each leaf the climate conditions at its canopy position, computes photosynthesis, and redistributes carbon for maintenance and growth based on age, size, and local environment.

That leaf-level state is then turned into an evolving 3D plant in NVIDIA Isaac Sim, where ray tracing estimates how much radiation each leaf intercepts and feeds that back into the growth model. For developers, the notable part is the bidirectional loop: structure affects light capture, and light capture affects structure over time, which is exactly the kind of coupling that makes digital twins useful rather than decorative.

Against greenhouse measurements, the model reported relative RMSE of 9.5% for total dry weight, plus 9.2% for leaf number, 12.7% for canopy diameter, and 13.1% for largest-leaf area. The paper also tested scenario changes: a 30% drop in incident radiation reduced final dry weight by 10.4%, a 30% increase raised it by 6.9%, and adding 200 ppm CO2 increased it by 46.1%.

The broader implication is that greenhouse control systems can reason about local microclimates instead of treating a crop block as uniform. In a simulated 40-plant block, interior plants ended up with 8.6% less dry weight than border plants, and the leaf-specific tipburn index rose in enclosed leaves as tipburn appeared in the greenhouse plants.