Observations along a transect, binned by depth

Long/tidy, one row per (station, depth bin, variable) — the shape both the matrix builder and the climatology consume.

Usage

cc_transect_section(
  con,
  line,
  cruise_key = NULL,
  variables = "temperature_ave",
  dataset_key = "calcofi_ctd-cast",
  depth_max = 500,
  depth_bin = 10,
  x = c("occupied", "line")
)

Arguments

con

DuckDB connection to a release.

line, cruise_key, dataset_key

as in cc_transect_stations().

variables

measurement_types to return.

depth_max

deepest bin, m (default 500 — a handful of casts reach 5000 m, and letting them set the axis squashes every standard cast into the top tenth of the plot).

depth_bin

bin width, m (default 10; floor bins).

x

passed to cc_transect_stations().

Value

Tibble: cruise_key, sta, dist_km, depth_m, variable, value.

Details

Depth is binned because CTD sensors sample continuously (47.283 m, 47.916 m, …), so grouping by exact depth deduplicates almost nothing and the native-resolution profile is jagged with sensor precision rather than signal. Bins are floor bins labelled by their shallow edge (floor(depth / bin) * bin: 0, 10, 20 …), the release's depth_bin convention and the grain of its climatology table, so cc_anomaly() joins exactly. 10 m, not 5: the released CTD series is thinned (a 10 m grid plus the profile's inflection points), so finer bins are built only from where the profile happens to bend.