After Saturday’s 44 mm storm, two of my non-vented level loggers were about 2 cm off the staff. In the field I go step-by-step: pause logging and note time, rinse off silt, two-point bucket check (ice bath about 0°C and warm tap about 35°C) to confirm span, resync clock to GPS time, apply baro correction from the nearby baro logger, record a manual staff reading, then restart and redeploy at the same elevation. Anyone tightening this workflow — do you add a third point or a soak to catch hysteresis in muddy water?
I’ve seen the same about 2 cm drift after a 40–50 mm squall — drives me nuts — and it turned out my baro logger was about 10–15 m higher than the stream, so the ‘apply baro correction’ baked in about 1+ cm bias. If your baro isn’t at the same elevation, a quick rule is about 1.2 cm water per 10 m elevation difference; I now park the baro at sensor height or add that offset in post. Also on your ‘two-point bucket check’, make the ice bath truly slushy at 0°C or it’ll read a bit high.
Quick example: after a 60 mm event I chased a 2 cm offset and it turned out the bracket had crept on the post, not the sensor. I put a Sharpie “witness mark” on the mount and do a quick “burp” with a syringe on the non-vented port to clear any bubble, which has saved me a few times. Small caveat: ice baths can read high if it’s mostly ice — make it a slushy mix so it sits near 0°C.