I just had this exchange with Decagon Devices -- the manufacturer of our soil moisture probes. In short, iron content will affect the probe response, so our gravimetric sampling is needed.
Hi Steve,
Whoa! Unusual minerology can have some strange affects on our probes. The probes emit an electromagnetic field that primarily detects water in soils, but iron is definitely going to affect the electrical properties of the soil. I think that you should be able to account for this with a custom calibration (http://www.decagon.com/education/calibrating-ech2o-soil-moisture-sensors-13393-04-an/). We have seen some strange things like this in volcanic soils. It isn't easy to tell what the exact effect it will have on a given soil, so it is definitely best to calibrate. I have cc'd Doug, one of our R&D scientists, who may have more to add on this topic.
Regards,
Chris Chambers
Name* |
Steven Oncley |
|---|---|
Company* |
NCAR/EOL |
Email* |
|
Country |
United States |
Phone Number |
|
How can we help you? |
We now have several EC-5 probes in and around the Arizona Meteor Crater. We are seeing large shifts from the default calibration (offsets of -2 to +10 %vol). I don't believe that this soil is very saline, however it does have a large iron content. (A magnet picks up dirt!) Does it make sense to you that iron in the soil would affect the EC-5 calibration? If so, would you expect an offset, gain, or other calibration shift? |
--
We Measure the World
==================
Customer Support Team
Decagon Devices, Inc.
2365 NE Hopkins Court
Pullman, WA 99163509-332-5600
509-332-5158 (Fax)
support@decagon.com
10/22/13
Summary:
- IOP5 is ending. Not ideal, since north winds were present much of the night.
Actions past 24 hours:
- near SPN1 fixed(?) by removing battery box diode
- flr SPN1 perhaps better after reorientation (shouldn't have been the case)
- rim/near guy wire tensions measured -- range 280--510
- gravimetric soil samples taken everywhere, offsets of -10 to +2% found. Data look good with this offset correction.
- NNE antenna connections checked and found to be okay. Perhaps a bit of blocking by solar panel.
- Almost all radiometers cleaned
To dos:
- Keep checking flr SPN1
- Generator servicing today
Sensor status:
T/RH: ok.
P: ok
csat u,v: ok
csat ldiag: ok
csat w, tc: ok
kh2o: ok. Again, one small glitch on flr yesterday.
motes: ok.
Wetness: ok.
radiation: ok. dfs.flr pattern a little different this morning.
Tsoil: ok
Gsoil: ok
Qsoil: ok. Offsets based on yesterday's gravimetric samples make the data look reasonable (for the entire experiment).
Cvsoil: ok.
2D sonic: ok
A quick in-and-out this afternoon.
Arrived at flr about 15:00
- rotated SPN1 120 degrees (slot in "band" is now east-west) to see if this changes the morning diffuse drop
- took soil sample
- checked NNE antenna. connections tight, antenna a bit blocked by solar panel, but I wouldn't expect this to be a big effect
- cleaned Rsw.in
Started up the trail at 15:50.
Steve S&O:
Far at 11:08.
Near at 11:47.
Flr at 15:35 (just Rsw.in, with some bird doo on the north side of the dome)
TWH, Nov 5: The date(s) on this are ambiguous. I think that these cleanings occurred on October 21, based on the date of the 5th IOP.
Samples taken today by Steve O&S:
To be expedient (and since not much structure in sandy soil), just took one sample 0-6cm.
At far, had to dig out core from the side and support from underneath to remove.
Despite almost 2 hours before weighing, saw no evidence of moisture condensed in container.
Site | Time | Weigh Time | Tare | "Wet" | Dry | Dry Density | Moisture (%mass) | Moisture (%vol) | EC5 reading (%vol) |
|---|---|---|---|---|---|---|---|---|---|
far | 11:20 | 13:10 | 3.1 | 176.7 | 172.2 | 1.23 | 2.7 | 3.3 | -1.2 |
near | 12:00 | 13:15 | 2.4 | 161.1 | 156.9 | 1.12 | 2.7 | 3.0 | 5.4 |
flr | 15:15 | 16:45 | 2.4 | 167.0 | 159.2 | 1.14 | 5.0 | 5.7 | -3.9 |
Treating the differences as a simple offset in the EC5 calibration, the Qsoil values for the entire project become somewhat believable.
We should contact Decagon to see if odd behavior in metallic soils would be expected...
P.S.(Mar 2014): During teardown, sensor depths were noted as 5, 3, 2 cm for near, far, flr, respectively, not just our usual 5cm at all sites. Since I only sampled 0–6 cm here, we can't look for any depth variation. However, we could expect that far and flr would read drier than near, as did occur. I don't know if 2 and 3cm are shallow enough to have suffered from lack-of-depth effects.
10/21/13
Summary:
- Only issue is both SPN1s.
- IOP5 will start tonight.
Actions past 24 hours:
- Reseated connector to revive SPN1.near (for second time)
To dos:
- Take gravimetric soil samples (now have all necessary equipment)
- Measure guy wire tensions
Sensor status:
T/RH: ok. RH.20m.rim is high by about 3%; RH.40m.near is high by about 1%. Not worth chasing.
P: ok
csat u,v: ok
csat ldiag: ok
csat w, tc: ok
kh2o: ok. One event about 1630 yesterday -- a bug?
motes: ok. SPN1.near died last night (just like the previous night), but revived by cycling power.
Wetness: ok.
radiation: Rsw.in MUCH better after cleaning; still odd SPN1.flr midday and at local sunrise
Tsoil: ok
Gsoil: ok
Qsoil: Still need to recalibrate far and flr.
Cvsoil: ok. Fixed range issue on Vpile.off.
2D sonic: ok
About 21:30, lost SPN1 data from near -- about the same time as the outage last night. Hard to believe that the connector slipped again at just this time. Vmote had decreased the entire day, but still was 12.4V at last report. Unlike last night, the mote isn't sending reset messages. Like last night, it doesn't respond to commands. Guess we'll take a look at it first thing tomorrow morning (about 9AM).
Using John's wisardMessageDecoder, I was able to see that Vpile.off values were larger than I thought. I've now (again) expanded the range to +/-50uV and now see data for these values and Lambdasoil (as expected). dsm server, etc was restarted.
Gordon: I now see 4 statsproc processes. This seems like one too many?
During last night's IOP, the users were having to cycle power to the Netgear WiFi box every 15min because of dropped connections. As a test, I have been running today using WiFi for more than 2 hours straight. Nevertheless, I just decided to change the WPA "key lifetime" from 60 -> 120 minutes. Perhaps this will keep the connection going longer?
Obviously, the other logical conclusion is that RF interference from things like the tethersonde that only operate during IOPs might be the source of poor connections?
10/20/13
Summary:
- Everything is working
- IOP4 ended successfully. Warm air intrusions occurred between 10pm and 2am.
Actions past 24 hours:
- Floor: cleaned radiometers (they needed it) and changed P.NNE battery
- Reseated connector to revive SPN1.near.
To dos:
- Take gravimetric soil samples (scales and oven should arrive today with Semmer)
Sensor status:
T/RH: ok
P: ok
csat u,v: ok
csat ldiag: ok
csat w, tc: ok
kh2o: ok. Sebastian cleaned flr yesterday.
motes: ok. SPN1.near died last night, but revived by reseating sensor cable.
Wetness: ok.
radiation: flr cleaned yesterday (and were dirty). Will watch to see if SPN1/psp comparison improves.
Tsoil: ok
Gsoil: ok
Qsoil: Still need to recalibrate far and flr.
Cvsoil: ok
2D sonic: ok
About 9:30 this morning (10/20), Sebastian changed the battery at NNE, in an attempt to solve the (somewhat) erratic data reporting from this barometer. Now I see min/max dTs of 0.73/1.31, which is somewhat better though still a larger range than any of the other crater barometers. Sebastian did note that there was a bit of water in one of the power connectors.
We have not seen a big outage since the event 3? days ago, which caused us to pay attention to this sensor.
About 20:00 last night, SPN1.near died. The mote was in a mode of resetting every 10 seconds, which didn't change in response to "rxb" or "xs" commands.
~10:45 Found that the DIN connector to the SPN1 was slightly pulled out. Reseated, cycled power on mote, and now works. Also measured voltage to SPN1 at 17.1V -- should be high enough! (I note that logged voltages are steadily dropping -- now at 15.3.)
Got this email from Sebastian this evening (19 Oct). See below that these cleanings were good.
Steve,
I cleaned the radiometers and krypton - there was lots of bird droppings on the upfacing pyranometer. It took 3 kim wipes to clean it all up ...
I started at 15:45 with the radiometers and finished by 15:55 with the krypton. But looking at the krypton plot now I think I made it dirty instead of clean! Maybe I should give it another try in the morning?
It's dark now, so I hope NNE will last through the night. I'll go check on it tomorrow morning, or if it drops out at night.
Have a good night, too!
-Sebastian
10/20: Effects:
- kh2oV 1.8 -> 2.2V
- Rsw.in: drop(!) of about 8 W/m^2. (P.S. Some morning and mid-day reductions in Rsw.in are now gone. This appeared to have started between 12--14 Oct -- all at the same time. This is consistent with Sebastian's comment about bird droppings, that would have shaded the sensor for high solar elevation angles.)
- Rpile.in: no effect
- Rsw.out: drop of perhaps 2 W/m^2
- Rpile.out: no effect
- SPN1: so far, better agreement between Rsw.global.flr and Rsw.in.flr (though global does go higher at about noon) and between Rsw.dfs.flr and Rsw.dfs.near (though not much data, since I just restarted near).
Filled both tanks on the ISFS generator today and the next time it needs fuel should be 22 Oct 13.
10/19/13
Summary:
- Everything is working!
- IOP4 begins tonight. Last night was a pretty good case (an "EOP" according to Sebastian), but tonight is supposed to be better.
- ISFS shift change: Tom flies out today; Semmer drives in tomorrow; I'm here
- ISS shift change: Cohn arrived yesterday; Phil leaves tomorrow.
Actions past 24 hours:
- Swapped TRH.30m.near. Temperature profiles have been improved and Ifan is normal.
- Reseated Xbee radio in soil.near. No data outages since.
- Tsoil.4.4cm.near appears to have been fixed magically by the soil mote work!
To dos:
- Swap battery on P.nne and/or check connections.
- Clean flr radiometers
- Take gravimetric soil samples (waiting for scales and oven)
Sensor status:
T/RH: ok. Down to -5 at the bottom of the crater last night.
P: ok, though data_stats shows P.nne with wider min/max sample reporting times than other crater P's.
csat u,v: ok
csat ldiag: ok
csat w, tc: ok. Sustained downdrafts at rim during 2 periods last night
kh2o: ok. An odd shift at near about noon (when we were there).
motes: ok!!! no outages.
Wetness: You've got to be kidding! Dry as a bone for days here... (ok)
radiation: still seeing odd SPN1 data, especially at flr. Those radiometers have never been cleaned -- will do today.
Tsoil: ok
Gsoil: ok
Qsoil: far and flr have been negative except for during the one rain event. Should be recalibrated with gravimetric (or assumed =0!)
Cvsoil: ok (I actually did this calculation yesterday. There was a shift after the rain 1.5 weeks ago, presumably as the probe settled into the soil.)
2D sonic: ok