My soil temperature experiment has been running for over 2 years. It has been a long time since I saw anything new in the data, so I am not too worried about it any more, but I figured that every years temperature profile is different, so I might as well let the experiment keep on running until I build…
Last time I downloaded the data, I noticed that there was one bad sensor, but I was busy and since I didn’t really need the data, I didn’t worry about it.
Several of the wires had been damaged, probably by little teeth… This white wire is the biggest sensor on the station and includes air temp and relative humidity.
Prepping for this next winter, I went out to the property today to change the batteries and download the data. I also replace the burlap skirt that I had over the wires. I remembered that one of the sensors had given bad data last time, so I took a careful look at the wires and found that two had be severed and a couple more were chewed partway thru.
Could this have been prevented? Possibly. In some of the pics, you can see what is left of the original burlap that I had used to protect the cords. It had been worn away by the environment (sun and weather), and I had been too slow to replace it. Had I replaced it sooner, I might have prevented whoever was chewing on the wires from having access. “Stitch in time saves nine.”
I downloaded the data and took a closer look… Things were a bit strange (as the charts will show), so I decided to try the number-one IT trick “turn it off and turn it on again”. I changed the batteries and rebooted the data logger.
The black wire was for the 7.5 ft deep control temp. Behind it, a bit more out of focus, you can see another black wire that was nicked…
Back at home this afternoon, I see that the data showed the 7.5ft deep control temperature probe (black wire) failed on June 6th, 2013 at around 5:30 PM. Instead of turning off and not giving any data, the sensor showed ridiculous numbers in the -90°C range.
In the middle of the night on September 6th, three months later, the air temperature and RH data (the white wire) went horribly wrong. Suddenly air temps appeared to drop from 12°C down to -91°C and relative humidity went from 93.9% to a nonsensical, but consistent, value of “1”. I don’t know why, but the air temperature fluctuated, so I guess some electrons were getting thru.
This chart shows the sudden drop in the air temperature data, along with subsequent fluctuations.
The data started getting strange lately…
Then something else strange happened… Shown as a blue vertical spike in the previous chart Not sure what, but on Oct 10th, the data suddenly went crazy… Perhaps the data logger hiccuped? Or the battery failed for a moment? Or maybe a deer peed on the station? I probably will never know, but this next chart shows the raw data around that time.
Raw data from my sensors showing the strange hiccup…
The data goes crazy for about 8 readings, but then things seem to reset. After the reset, everything seems to have shifted down a bit… The strangest thing is that the 7.5 ft deep temperature sensor and the air temperature and RH sensors started making sense again (see columns H, L and M, before and after the hiccup). I can’t think why the data would start making sense with the wire still cut… perhaps there was data in the cache and it is showing me samples from a previous year? Anyway, when I saw this, I was glad I had reset the system while I was out there today… Hopefully things will look better next time I download the data.
This next chart shows all the temperature sensors, but I removed the 8 really bad readings from the above chart… You can still see the sudden shift on October 5th, as well as how long the 7.5 ft deep gauge was misbehaving… You can also see that everything returned to “normal”, but shifted downward a bit.
Fortunately, I had already determined that there was hardly any difference between the dry and control temperatures at 7.5 ft deep, so I won’t miss that sensor too much… But I will need to get out there and repair the other sensor when I get a chance and see if the probes appear to be trustworthy again after their reset.