Repair and re-derive a table of BioSIM FWI_Daily records

The single entry point for making BioSIM FWI_Daily output trustworthy. Used by get_fwi_daily(), and exported so that projects with their own BioSIM fetch apply exactly the same correction rather than reimplementing it.

Usage

repair_fwi_daily(fwi_daily, validate = TRUE)

Arguments

fwi_daily

data frame of FWI_Daily records with at least FFMC, DMC, DC and WS.

validate

logical. Check the repaired values against physical bounds and error on a violation. Default TRUE; a violation means a failure mode the repair does not characterise, which must not reach a weather database silently.

Value

fwi_daily with FFMC, DMC and DC repaired and BUI, ISI and FWI re-derived.

Details

Two steps:

1. Repair the three moisture codes FFMC, DMC and DC in place, via repair_fwi_exponent(). These are recursive, precipitation-driven daily codes, and FWI_Daily does not return precipitation in every configuration, so they cannot in general be recomputed from the other columns and must be recovered as they stand. Repairing DMC and DC matters as much as FFMC: a corrupt DMC otherwise flows into the buildup index and from there into FWI.

2. Discard and re-derive the three behaviour indices BUI, ISI and FWI from the repaired codes using cffdrs. These carry the same artifact as the codes – BUI drops below 1e-4 whenever DMC does – but they are pure functions of the codes, so recomputing is strictly safer than repairing: it needs no threshold and cannot mistake a genuine extreme for corruption. Nothing is lost by doing so. On uncorrupted records the recomputed values reproduce BioSIM's own to within 5e-6 relative for BUI and 0.8% for ISI and FWI.

Records BioSIM did not model (-9999 sentinels, which callers map to NA before calling) pass through as NA.

See also

repair_fwi_exponent(), get_fwi_daily()