FAQ

Do I need SIMPLACE installed to use simpest?

Only for the crop growth stage (simpest.models.simplace). The disease/fungicide simulation stage (simpest.models.franchestyn) takes plain pandas DataFrames as input, so you can run it on its own — using crop-model output produced elsewhere — without a SIMPLACE installation. See Installation.

Why is my disease severity always zero?

Check that disease_type is set on FranchestynConfig (or the runner's disease_type argument). Passing None intentionally disables the epidemiological model and gives a crop-only run. Also confirm the crop-model series covers the sowing-to-maturity window: the disease step only runs on days after sowing.

My calibration and simulation results don't match a reference run

Two settings commonly explain a mismatch:

• use_gdd. If the reference crop-model series ties its own cycle completion to accumulated thermal time rather than a fixed calendar length, set use_gdd=True on FranchestynConfig so simpest's cycle completion is computed the same way. Leaving it at the calendar-day default in that case makes the two cycle percentages drift apart over the season.
• use_prev_day_alignment. This controls whether the calibration objective compares each simulated day to the reference on the same date or the next date (sim[d-1] vs ref[d], the default). Confirm it matches the convention used by whatever reference implementation you're comparing against.

Calibration is slow — how do I speed it up?

Reduce n_restarts and/or max_iter on FranchestynConfig, narrow calibration_variable to "crop" or "disease" instead of "all", or exclude parameters you don't want searched with crop_disabled_params / disease_disabled_params (or deactivate_calibration). Fewer free parameters and fewer restarts both reduce the number of model evaluations.

Why did a fungicide treatment have no effect?

fungicide_type must be set (not None) for the fungicide step to run at all, and the management data must include a scheduled treatment date. A missing treatment date is a silent no-op rather than an error, so check the management CSV's treatment columns if efficacy stays at zero.

Where do the crop, disease, and fungicide parameter defaults come from?

From the JSON files bundled with the package (simpest/models/fr_crop_parameters.json, fr_disease_parameters.json, fr_fungicide_parameters.json), keyed by crop/disease/fungicide type. Each parameter carries a default value, calibration bounds, and a calibration flag; see Usage for how to override or exclude them.