I deployed one of the prototypes in our local tidal estuary over a 24-hour period to assess the behavior of the instrument. Readings were validated against commercial sensors and found to be within 2% accuracy.
A few interesting things to note: The weird "shakes" at the beginning, and then twice near the end of the series, are me, knocking the device. A few hours after deployment I checked it for leaks, and then later, near the end of deployment, I knocked it a few times to check if it had filled with water. The housing and electronics remained completely dry.
The steep drops or rises appear to be caused by the device sliding up and down on the rebar stake. Right now the instrument is slightly buoyant, and when the tide comes in, it lifts the device with it. Future deployments will use a more robust locking system to keep the CTD in place.