Mean-field consistency, not biological calibration

Cropmix 0.2 deliberately separates biological dispersal information from mean-field consistency. The function assess_mean_field_consistency() scans a movement-scale grid and asks at which scales the finite spatial model becomes practically indistinguishable from the PLOS-style deterministic mean-field reference.

The returned one-standard-error scale is therefore a mean-field adequacy threshold, not an estimate of insect dispersal. Biological analyses should use a scale supported independently by the literature, tracking data, or spatial epidemic observations.

The deprecated name calibrate_kernel() remains as a compatibility alias.

result = cropmix.assess_mean_field_consistency(
    field, system, scenarios, reference_variety="SUSC"
)

For the CBSD manuscript case study the primary spatial scale is literature based. McQuaid et al. (2016) use an exponential radial kernel with mean dispersal distance 2/alpha = 5.25 m (range 1-11.5 m). With Cropmix's kernel K(d)=exp(-d/a), this corresponds to a = 2.625 m and a sensitivity interval a = 0.5-5.75 m before finite-field balancing.