The channel width link reads on a stream nobody measured, and the waterbody tests

scripts/0296's width. link joins whse_basemapping.fwa_stream_networks_channel_width province-wide, and its FIELD_MEASURMENT rows are FDIS sites, so at a visit the stored width can be that visit's own measurement, which link never sees on the unmeasured streams it classifies. So fwapg's own width model (upstream area and upstream mean annual precipitation, Thorley and Irvine 2021b) is recomputed here for the segments' watershed codes, and cw_deploy is the width link would read with no field measurement: the stored width unless it is FIELD_MEASURMENT, then the model where stream_order > 1 (fwapg models no order-1 stream) and NA otherwise. The recomputed model must match the stored MODELLED widths within tol (relative) on at least agree_min of them, or the call stops: the stored widths come from older inputs, and a larger disagreement means the formula or its inputs have changed.

Usage

knw_link_width(
  conn,
  linear_feature_id,
  tol = 0.05,
  agree_min = 0.95,
  wetland_ha_min = 1,
  lake_ha_min = 100
)

Arguments

conn: a DBI connection to a database holding whse_basemapping.
linear_feature_id: numeric FWA segment ids.
tol, agree_min: numeric(1) the model-agreement control.
wetland_ha_min, lake_ha_min: numeric(1) link default's area floors.

Value

A list: widths, one row per segment (linear_feature_id, fwa_waterbody_key, fwa_channel_width, fwa_cw_source, fwa_stream_order, cw_model, cw_deploy, wb_river, wb_wetland, wb_lake, wb_any_lake); control, a one-row data.frame with n_modelled, agree_share and median_ratio.

Details

Alongside the width, the waterbody tests of link's default rules: whether the segment's waterbody_key is a river polygon, a wetland of at least wetland_ha_min, or a lake or man-made waterbody of at least lake_ha_min.

Examples

if (FALSE) { # \dontrun{
conn <- DBI::dbConnect(RPostgres::Postgres())
knw_link_width(conn, c(701747495, 701747496))$widths
} # }