Skip to contents

Choose the best diagnostic engine for a single pedon

Source: R/engine-selection-v0965.R
pick_engine.Rd

Per-pedon heuristic: returns "aqp" if the pedon's horizon table has the morphological richness that makes aqp's canonical NRCS dispatch reliable, otherwise returns "soilkey" (the more permissive hand-coded path).

Usage

pick_engine(pedon, min_score = 3L)

Arguments

pedon

A PedonRecord.

min_score

Integer (1-5). Minimum completeness score for "aqp" engine to fire (default 3).

Value

Character: "aqp" or "soilkey".

Heuristic

We score each pedon on a 0-5 morphology-completeness scale; aqp fires when score \>= min_score (default 3). The five axes:

  1. Designation present (any layer has a non-blank designation, e.g. "A1", "Bt2", "Bw").

  2. Texture quantitative (any layer has both clay_pct and sand_pct populated).

  3. Munsell complete (any layer has all three of munsell_hue_moist, munsell_value_moist, munsell_chroma_moist populated).

  4. Structure recorded (any layer has a non-blank structure_grade).

  5. Clay films / argic evidence (any layer has a non-blank clay_films_amount or designation pattern matching Bt).

Why this matters

On BDsolos RJ (data-rich), the heuristic recommends aqp for ~99 canonical thresholds). On LUCAS topsoil-only (data-sparse), it recommends aqp for ~0 clay-films / designation axes are unfilled. Calling classify_*(pedon) routed through the heuristic gives the correct engine per pedon, recovering both the BDsolos RJ lift AND the LUCAS robustness.

See also

argic, cambic.