Roadmap​#

PlantSimEngine uses the same simulation engine for a single object or many plants and organs, including models that run at different time steps. This page lists work planned to extend and check these capabilities.

Current priorities are:

• migrate downstream model packages to CompositeModel, CompositeModelTemplate, ObjectInstance, and ModelSpec;

• strengthen type-stability and allocation tests for million-object workloads;

• test more combinations of adding, removing, and moving objects, including updates to their local growing conditions;

• make error reports clearer when object selections match too many objects, models try to set the same output, or time-step connections need attention;

• validate mutable voxel, layer, and octree microclimate backends;

• test more dependent packages and simulation performance before releases;

• investigate running independent groups of model calculations in parallel.

Environment and microclimate work​#

Trial environment sampling​#

Coupled microclimate solvers can iterate on local environmental state before accepting a timestep. A canopy energy-balance model, for example, may need to:

1. propose a trial canopy air temperature and humidity;

2. run leaf models against that trial air state;

3. update the trial air state from leaf sensible and latent heat fluxes;

4. repeat until convergence;

5. commit only the accepted canopy or voxel air state to the mutable environment backend.

Pass non-committing trial state through run_call!:

run_call!(context, :leaf_energy; environment=trial_environment, publish=false)

Then commit the accepted state through the model-facing environment API:

commit_environment!(context, accepted_environment)
run_call!(context, :leaf_energy; publish=true)

Each leaf still receives conditions for its own location, so one call can supply different trial values to different leaves. commit_environment! saves only the accepted growing conditions. Future work will test this approach with environments represented by cells, layers, and octrees.

The full issue list is available on GitHub.