Depth-selective withdrawal (hypolimnetic intake). Supported?

[fpacheco]

I'm migrating from EFDC to SCHISM, and this is my first WQ reservoir application - so apologies if I'm
missing something obvious.

I'm modeling a small, seasonally-stratified reservoir (baroclinic run, ICM water quality) and need
to represent a real bottom water intake: a submerged valve at a fixed invert elevation (~20.6 m,
well below the surface, in the hypolimnion), used to draw cold water to downstream.

I looked for a way to make either source_sink (vsink.th) or a hydraulics.in structure
(orifice/pipe, which does take an elev/height) withdraw water selectively from that vertical
band, rather than depth-uniformly across the whole local water column.

Before asking, I checked:

1. The main manual (v5.9, §5.17): "the point source approach injects flow via the continuity
   equation alone without providing extra momentum" - so source_sink is depth-integrated, as
   expected.
2. The dedicated hydraulics/structures manual (structs_main.pdf): elev/width/height for
   orifice/pipe/radial only parametrize the stage-discharge relationship (how much flow, as a
   function of head above elev, capped by height) - I don't see anything about where in the
   water column that flow is drawn from.
3. The source (schism_step.F90, hydraulic_structures.F90): the cross-sectional area used to turn
   q_block into a velocity (vnth_block = q_block/ar) is computed from the FULL local water
   column height at the reference side (htot = max(h0, dps+eta_avg)), independent of the
   structure's elev/height. That velocity is then applied uniformly to every vertical level at
   that face/element - so even an "orifice" positioned at depth ends up withdrawing uniformly from
   the whole column, not selectively from the band near elev.

Question: is there an existing mechanism (maybe undocumented, or planned/in a dev branch) to
achieve genuine depth-selective withdrawal - i.e., remove water (and its tracer concentrations)
preferentially from a specified elevation band, mimicking a real selective/hypolimnetic intake in a
stratified waterbody? Or is this a known, accepted limitation of the source_sink/hydraulics
modules?

If this is a limitation, is there a recommended alternative solution for simulating deep water withdrawal from reservoirs?
Thanks in advance!