Relative geostrophic velocity between adjacent CTD stations

Rasmus Swalethorp's recipe (email 2026-09-22; CalCOFI/workflows#103): per station, absolute salinity and conservative temperature on a dp dbar grid to p_ref, dynamic height anomaly relative to p_ref (gsw_geo_strf_dyn_height()), then for each adjacent pair v = (dh2 - dh1) / (f * dx) with f the Coriolis parameter at the pair's mean latitude and dx the distance between the two stations. Positive is 90 degrees to the left of the direction of increasing station order — for stations ordered nearshore to offshore on a CalCOFI line (which runs west-south-west) that is equatorward.

Usage

ctd_geostrophic(casts, p_ref = 500, dp = 1, min_dx_km = 10)

Arguments

casts

data frame, one row per sample, ordered or orderable by station_order: columns station (id), latitude, longitude, pressure (dbar), temperature (deg C), salinity (PSS-78); optional q_temperature, q_salinity, station_order (numeric, default the order of first appearance) and dist_km (along-line distance; default the great-circle distance between the pair).

p_ref

reference pressure (dbar), default 500.

dp

grid spacing (dbar), default 1.

min_dx_km

minimum station spacing (km), default 10. A station closer than this to the last station kept is skipped: the geostrophic shear is 1 / dx, so the extra inshore stations a few km apart (the SCCOOS 90.27.7 beside 90.28, 93.26.4 beside 93.26.7) turn a small density difference into metres per second of spurious flow.

Value

tibble, one row per (station pair x pressure): station_1, station_2, dist_mid_km (from the first station), dx_km, pressure, velocity_m_s, shallow.

Details

The flow is relative to p_ref (there is no level of known motion), so it has no anomaly. A station shallower than p_ref has its deepest density carried down (approx(rule = 2), as in the recipe) and the pair is marked shallow = TRUE. Flagged temperature or salinity samples are dropped first; a station with fewer than 2 good samples is skipped.