�From:�	Andrew Gray (graya)
�To:�		sowacs@aqua.ccwr.ac.za
�Date:�	Thursday, June 20, 1996  7:09 pm
�Subject:�	Re: TDR calibration

Bruce Meterlekamp wrote:
  >4/ Another point I am trying to get discussion on is
  >how to calibrate TDR (and related) sensors.  A simple two-point(oven-dry
  >and saturated) is surely not appropriate if the main use forthe sensor
  >will always be inbetween these. Obtaining uniform watercontents at points
  >in between will surely involve either a lot of mixing andrepacking of
  >sequntially wetted soil, or long time periods to allowredistribution of
  >soil water added from the top or bottom?

  I agree that a two-point calibration is probably notappropriate, because of
  the non-linear (1/V) response of TDR at low soil-moistures(especially
  oven-dry), but if your soils in the field never gets too dry(e.g. 12% or
  less), maybe a low and high calibration point would be enough?

  I am extremely skeptical of any calibration involving mixing andpacking of
  soils, particularly after sieving the soils to 2 mm.  Thatdestroys much of
  the soil structure, removing aggregates and air pockets (thelatter which
  may greatly influence the TDR response), so my feeling is thatyou're
  calibrating something unrelated to the field soil, particularlyin the
  high-organic content forest soils I work in.

  I conducted my calibration similar to the method in Herkelrathet al
  1991 (Water Resources Research 27: 857-864) and collected soilsamples by
  pounding 10 cm diameter PVC pipe into the ground and extractingthe core
  (with minimal compaction), letting the core dry in the lab, thensealing the
  bottom and adding water to the top and taking measurements.  Ididn't wait
  more than a few minutes before sampling, because TDR is robustto vertical
  heterogeneity in H2O content along the rods.  Even thoughhorizontal
  heterogeneity might have been a factor, the confidence intervalson my
  regressions were still around 2.5% H2O, which isn't much worsethan Topp et
  al's.

  Some folks I know have been doing point sample calibrations, bytaking TDR
  readings in the field at different times of the year, thenextracting a soil
  core at the sample point.  My concern there is that a standardsoil sample
  tool is only about 3 cm in diameter, and I doubt you can get anaccurate
  volume sample over a depth of 30 cm or more (soil compacts anddoesn't enter
  the bottom of the sampler).

  I'd really appreciate any feedback on these methods.  Although Ipublished
  my calibrations (Can. J. Forest Res. 25: 376-385), I haven'theard a peep!

  Andrew Gray
  graya@fsl.orst.edu