Do you really only need to know whether the water is up to a certain level, or are you ultimately trying to sense the level. If the latter, there are some technologies you might not be aware of. A company I work for specializes in measuring levels in tanks, and things like the draft of ships. We have off the shelf products, and some new ones in development, that can measure the level of rivers or the ocean.

Olin Lathrop‭ The purpose of the electrodes is to measure conductivity (salinity). I’d like to cover the range from 0.1 to 50 mSiemens/cm (surface water, and seawater). I can settle for 0.1 to 8 mSiemens/cm (surface water only). The device will be completely submerged all the time.

By the way, I’m going to use an LPS28 pressure sensor to measure depth. Correction for barometric pressure will be done during post-processing.

I think keeping the electrodes clean will be a problem. Gold should be inert enough, even in seawater. But the chlorine in the salt has a way of getting into everything eventually. Soldermask might survive long term in clean water, but I'd really want to do accelerated aging tests before expecting it to survive years in seawater.

I tried measuring water conductivity once using graphite electrodes. I thought they wouldn't corrode. They didn't, but the readings weren't repeatable. The project moved on, and I never did figure out what was going on there.

Olin Lathrop‭ I like the idea of an accelerated lifetime test. If I sous vide the sensor in seawater for 2 months at 70°C, that should be equivalent to a year in the field. (Assuming the common rule of thumb that every 10°C accelerates damages by a factor of 2x). I could freeze it a few times along the way too, for a good measure.