If it isn't making sense to you perhaps very gently bubble some gas through a hose sitting deep in water in a bucket. As your raise and lower the tip of the hose in the bucket of water the hose pressure will have a pressure that will equal the cm of water that sit above the bubbling end of the hose.
If the end of the hose is gently bubbling you'll get a reasonably accurate pressure (if the flow rate is too hi you'll start getting artifacts)
Back in the posts someone discovered the temperature had to be read from the sensor each time before each of the pressure readings or else the pressure readings would not be accurate. Maybe something related to that?