Skip to contents

Load the bundled KSSL + NASIS morphological-enriched sample (v0.9.75)

Source: R/wosis-cache.R
load_kssl_nasis_sample.Rd

Returns a 99-profile snapshot built by joining the NCSS Lab Data Mart (ncss_labdata.gpkg) with the companion NASIS Morphological sqlite (NASIS_Morphological_*.sqlite) via load_kssl_pedons_with_nasis. Pre-annotated with derived WRB Reference Soil Group via usda_to_wrb_rsg.

Usage

load_kssl_nasis_sample()

Value

A list with pedons, pulled_on, source, join_helper, cross_walk.

Details

Compared to load_kssl_sample (KSSL lab tables only), this sample carries the morphological evidence that several WRB diagnostic horizons need:

| Field | KSSL-only | KSSL + NASIS | |——-|———-:|————-:| | munsell_hue_moist | 0 | munsell_value_moist | 0 | munsell_chroma_moist | 0 | munsell_hue_dry | 0 | structure_grade | 0 | structure_type | 0 | clay_films_amount | 0 | slickensides | 0

First-ever benchmark on this enriched sample (soilKey v0.9.75, full v0.9.69-72 fallback stack):

  • Top-1 baseline: 19.1\ **+3.5pp lift purely from NASIS morphology**)

  • Top-1 full stack: 20.6\

  • Phaeozem: 1/33 -> 2/33 (Munsell-driven mollic detection)

  • Podzol: 0/15 -> 1/15

Remaining ceiling driven by attributes neither dataset preserves: Solonetz needs Na/ESP, Vertisols need slickensides + cracks (NASIS records 1.7 on subsoil samples NASIS often lacks.

Reference

Beaudette, D., Skovlin, J., Roecker, S., Brown, A. (2024). aqp: Algorithms for Quantitative Pedology. R package version 2.x. https://github.com/ncss-tech/aqp.

Examples

if (FALSE) { # \dontrun{
s <- load_kssl_nasis_sample()
length(s$pedons)
#> 99
# Munsell now populated (KSSL-only sample had 0%):
mean(vapply(s$pedons,
            function(p) any(!is.na(p$horizons$munsell_hue_moist)),
            logical(1)))
#> 0.99
} # }