TobyDJackson
commented
7 years ago So, are you interpreting that strain data as a sequence of slight changes (2.9-3.03) around a displaced position (mean or mode approx 2.97) i.e - the changes in strain are only ~3% of the total strain in that 10 minute segment. Or, will you subtract some offset from the strain data?
I find the same type of correlation. I then pick a mean wind speed and a maximum strain, assuming that subtracting the running mode will take away any offset due to temperature, equipment, tree growth etc. I believe this is the way it has previously been done in quite a few papers, I will ask Barry for some input here.
Hi Toby et al. Hope I do this correctly (not used to GitHub). To understand why we may have different perspectives, I attach a figure, showing the results from our very first day of observations. Wind speed is in blue and strain (voltage) is in red. I was immediately impressed by the close correlation also at high temporal resolution. Our goal is to find a unique relationship between recorded wind speed (and temperature) just upwind of the tree to the integrated drag of the tree. We are hoping for easily applicable temperature corrections (not sure if this will be the case) and that a few calibrations where we pull the tree with a known force will allow us to get to the drag.
I guess I don't quite understand your method. When you relate your fluctuation in strain to wind speed (where you have removed the mode of the strain), do you then pick a max value from the wind observations or the mean wind speed? For me, it would seem more intuitively correct for you to use the observed (max wind speed minus the observed mean wind speed) and relate those to the (strain minus mode/mean of the strain).