Some great points mikem & Bob Mehew
That ultra sonic cave logger looks awesome!
In terms of rainfall data I think you're right in that trusting a single source would likely lead to inacurate/misleading results depending on its location and mechanism by which water enters the cave. I think to begin with I'd cast the net quite wide around a cave in terms of rainfall data collection and see what the data looks like when compared to cave water levels over time.
This is definetly more a more complext probelm that I first thought and it's likely the model would need to have a large margin for error. Also taking local river levels into account sounds like a great idea too.
Here's a bunch of EA weather stations around Horton in Ribblesdale as an example.
Plots showing rainfall data from from the past week at all of the stations shown above.
Chocolate fireguard said:
This is a great idea.
For about 20 years I ran research projects at a couple of industrial research organisations.
One of the things I learned was that when you are first trialling an instrument you have built or modified you need to know what the results are going to be. If you don?t you have no idea if any surprises are due to genuine effects or your equipment not behaving as you expect.
I would not use either Peak or Sleets Gill as a first step. So far as I am aware no one (yet) understands how or why they respond to different sources of water, and monitoring those sources would not be straightforward.
In Derbyshire I might go for Giants, perhaps at sump 1. In Yorkshire perhaps Upper Long Churn. Both easily accessible and so far as I know the source of water in each case is obvious and rainfall data is available.
It is quite likely you already knew all this. If so please don?t be offended, but there has been no mention of it so far on the thread.
Good luck with it, and please do continue to post your progress.
I agree a pilot study on a cave with reasoably easy to understand/predictable hydrology would be a good place to start. I'm thinking somewhere that normally has welly deep water but occasionally floods, this way I might be work out signal drift over time as the water levels return to a baseline during dry periods I can do an aproximate recalibration, that is if the sensor remains precise but becomes inaccurate over time, something I'll have to look into. Alternativley I can leave a measuring stick in there and take manual water level readings for data calibration when I come to process the data.