computePunstable {sarp.snowprofile.pyface}

Compute probability of layer instability based on random forest model

Description

This function enables comfortable and fast R access to Stephanie Mayer's python implementation of her random forest model to estimate the probability of dry snow layer instability. The routine can be run very efficiently on large snowprofileSets. Layer properties required are sphericity, viscous deformation rate (10e-6 s-1), density (kg m-3), grain size (mm), and the critical crack length (m) (which can be computed very efficiently automatically if shear strength (kPA) is available.) Additionally, skier penetration depth in (m) is required.

Usage

computePunstable(x, ...)

## S3 method for class 'snowprofileSet'
computePunstable(
  x,
  ski_pen = NA,
  recompute_crit_cut_length = TRUE,
  buffer = TRUE,
  ...
)

## S3 method for class 'snowprofile'
computePunstable(x, ski_pen = NA, recompute_crit_cut_length = TRUE, ...)

## S3 method for class 'snowprofileLayers'
computePunstable(x, ski_pen = NA, ...)

Arguments

x

snowprofile, snowprofileSet, or snowprofileLayers

...

passed on to subsequent methods

ski_pen

skier penetration depth (m), one scalar for each profile in x

recompute_crit_cut_length

This routine can very efficiently compute the critical crack length with computeCritCutLength. SNOWPACK often provides NA values of the critical crack length even for layers that have a real solution to it. With this flag you can conveniently recompute all critical crack lengths (TRUE). If set to FALSE, it will only be computed if not all profiles already contain it. Note that shear strength must be available to compute the critical crack length!

buffer

internal switch to ensure fast computation at low memory cost. Leave at TRUE!

Value

x is returned with ⁠$p_unstable⁠ (and potentially ⁠$crit_cut_length⁠, ⁠$slab_rho⁠, and slab_rhogs) appended to each profile's layers object.

Methods (by class)

• snowprofileSet: for sarp.snowprofile::snowprofileSets

• snowprofile: for snowprofiles

• snowprofileLayers: for snowprofileLayers

Author(s)

fherla and smayer

References

Mayer, S., Herwijnen, A. Van, Techel, F., & Schweizer, J. (accepted, 2022). A random forest model to assess snow instability from simulated snow stratigraphy. The Cryosphere Discussions. https://doi.org/10.5194/tc-2022-34

Examples

## load a handful of example profiles from a PRO file
profiles <- snowprofilePro(system.file("extdata/snowprofile.pro",
                                       package = "sarp.snowprofile.pyface"),
                           remove_soil = TRUE, suppressWarnings = TRUE)
summary(profiles)
names(profiles[[1]]$layers)
## compute p_unstable alongside critical crack length, slab_rho, slab_rhogs:
if (have_dependencies()) {
profiles <- computePunstable(profiles)
names(profiles[[1]]$layers)
}