{# Start/end-windows explainer — the `windows_note` panel payload (kind="html"). Static prose + nested list; shown above the window figures. #}

Left panel = first 7 h (1 h lead-in + 6 h of record) | Right panel = last 7 h (6 h of record + 1 h tail). ││ orange dashed = auto-detected (pressure-based) suggested deploy/recover time • ││ green dashed = YAML deployment_time / recovery_time • ││ grey = stage 1 raw record (same variable, same units).

Reading the grey background: where the raw stage 1 record is available and uses the same physical units as the stage 2/3 data, it is shown in light grey so you can visually verify that the chosen deployment window captures the in-water period. Cases where grey may be absent:

Sensors without their own pressure (e.g. a microCAT that relies on interpolated pressure from a nearby instrument) — interpolated pressure is only added at stage 3, so stage 1 has no pressure variable and no grey trace appears.
Velocity in beam or XYZ coordinates — if an Aquadopp or ADCP was configured to record in BEAM or XYZ mode, the stage 1 file contains beam/XYZ velocity variables while stage 3 stores ENU velocities; the variable names differ so no grey velocity trace is shown.
Conductivity — raw conductivity units sometimes differ between stage 1 and stage 3 (e.g. mS/cm vs S/m) and the grey trace is suppressed automatically to avoid a misleading scale mismatch.

If the orange vline is absent from the right-hand panel, no pressure-based recovery transition was detected in the final 25 % of the stage 1 record — the suggested recovery time equals the last raw sample. Check the timing table in the mooring summary report for the suggested UTC time.