Skip to contents

Depth at which a profile reaches a threshold: the nitracline and hypoxic-boundary rule

Source: R/ctd_derived.R
profile_threshold_depth.Rd

Rasmus Swalethorp's definition (email "Next Two Weeks Tasks" 2026-09-24; the CalCOFI legacy D_NClineDepth.m): the nitracline is the "depth at which nitrate is equal to or just above 1 uM. For the 1 m binned data this is simple. For the discrete depth bottle data we will have to find the depth between the point above and below this concentration ... by linear interpolation between the two points". The hypoxic boundaries are "the same as nitracline" with oxygen falling to a threshold, taking the shallowest crossing (see ctd_hypoxic_depth()).

Usage

profile_threshold_depth(
  depth,
  value,
  qual = NA,
  threshold,
  direction = c("above", "below"),
  surface_max_m = 20,
  min_n = 3
)

Arguments

depth

depth (m), one cast or bottle profile.

value

the quantity tested (nitrate in umol/L, oxygen in mL/L, ...).

qual

quality codes of value (NA = good); 8/9 are dropped.

threshold

the concentration the profile must reach.

direction

"above": the first depth where value >= threshold (a nutricline); "below": the first depth where value <= threshold (a hypoxic boundary).

surface_max_m

deepest first sample that may be reported as the crossing when it already reaches the threshold (m), default 20.

min_n

fewest samples the profile needs for an at_surface answer, default 3.

Value

one-row tibble: depth_m, value_above, value_below (the bracketing values, the same value twice for at_surface), status (ok, at_surface, no_bracket, not_reached, no_data), threshold, direction, depth_max_m, n.

Details

The rule, as the legacy script applies it: the profile is cleaned (ctd_mld()'s rule: flagged 8/9, missing and non-finite samples dropped, duplicate depths averaged, sorted), then the shallowest sample i that reaches the threshold is found — value >= threshold for direction = "above" (nitrate; the script tests NO3 > 0.9999, so a tolerance of 1e-4 is allowed), value <= threshold for "below" (oxygen). If i > 1 the crossing depth is interpolated linearly between sample i - 1 and i (status = "ok"). If the first good sample already reaches the threshold, its depth is the answer only when it is no deeper than surface_max_m and the profile has at least min_n samples ("at_surface"; the script: "I_NO3 == 1 && Depth < 20 && NoDepths > 2"); otherwise there is nothing to bracket ("no_bracket"). A profile that never reaches the threshold is "not_reached"; fewer than two good samples is "no_data". Only the shallowest crossing is reported: an oxygen minimum in mid-profile has a lower boundary too, which this rule does not return (Rasmus Swalethorp, 2026-09-24: "I suggest we just focus on the shallowest boundary depths").

Examples

profile_threshold_depth(c(0, 10, 20, 30, 50), c(0.1, 0.2, 0.5, 1.5, 8), threshold = 1)   # 25 m
#> # A tibble: 1 × 8
#>   depth_m value_above value_below status threshold direction depth_max_m     n
#>     <dbl>       <dbl>       <dbl> <chr>      <dbl> <chr>           <dbl> <int>
#> 1      25         0.5         1.5 ok             1 above              50     5
profile_threshold_depth(c(0, 50, 100, 200), c(6, 4, 2, 1), threshold = 2.4, direction = "below")
#> # A tibble: 1 × 8
#>   depth_m value_above value_below status threshold direction depth_max_m     n
#>     <dbl>       <dbl>       <dbl> <chr>      <dbl> <chr>           <dbl> <int>
#> 1      90           4           2 ok           2.4 below             200     4