Diagnose a variable-aware vertical transition-layer candidate

Interpret a generic C5 gradient candidate as an unthresholded thermocline, halocline, branch-aware oxycline, or C9 potential-density pycnocline candidate only when certified semantic identity authorizes that interpretation.

Usage

transition_layer(
  x,
  diagnostic,
  variable = NULL,
  support = c("local", "all")
)

Arguments

x: A direct certified source-profile <ocean_cube>, an already computed certified C4 vertical-gradient <ocean_cube>, or a certified C9 thermodynamic state for the pycnocline route.
diagnostic: Required diagnostic name: "thermocline", "halocline", "upper_oxycline", "lower_oxycline", or "pycnocline".
variable: Optional exact current variable name. NULL selects the sole variable whose preserved source CF standard_name is eligible; zero or multiple eligible variables are errors.
support: C5 support policy. "local" excludes explicitly gapped secants; "all" may return a truthfully labelled gapped candidate.

Value

A base data.frame with one row per longitude, latitude, and time. C5 feature diagnostics are retained and augmented with physical-variable identity, diagnostic definition, basis, direction, strength, input mode, diagnostic status, Provenance V1, and bounded QA.

Details

Physical eligibility is determined only from immutable preserved source CF standard_name and a bounded compatible-unit contract. Variable names and long_name are never semantic evidence. Thermocline selects the strongest eligible negative temperature gradient with depth. Halocline selects the strongest eligible absolute salinity gradient and preserves its sign.

Direct source profiles compose depth_gradient() and depth_feature(). Existing certified gradients are not differentiated a second time. Gaps, ties, missingness, completeness, source brackets, spacing, support gaps, and certification limits remain authoritative. The pycnocline route accepts only C9 potential density referenced to 0 dbar and uses the strongest positive gradient; it is an operational candidate rather than a universal complete layer definition. Oxygen diagnostics first resolve a complete-profile minimum/core, then rank negative gradients on the upper branch or positive gradients on the lower branch. No threshold, smoothing, prominence, width, curvature, or mixed-layer depth is calculated. localization_half_span_m is a vertical resolution scale, not statistical uncertainty.

Examples

if (FALSE) { # \dontrun{
transition_layer(cube, diagnostic = "thermocline")
transition_layer(cube, diagnostic = "halocline", support = "all")
transition_layer(cube, diagnostic = "upper_oxycline")
transition_layer(cube, diagnostic = "lower_oxycline")
transition_layer(state, diagnostic = "pycnocline")

gradient <- depth_gradient(cube)
transition_layer(gradient, diagnostic = "thermocline")
} # }