proudly sponsored by
|
archiving of SOWACS proudly sponsored by |
|
to all, i have also have been a lurker on the sowacs listserver for quite some time. but the current discussions, from grant upchurch, steve evett and others, regarding TDR calibration have been my interest, so here goes. 1. recently, i worked on the development of a transient procedure for calibrating TDR probes/systems. the reason for the development was the fact that steady-state methods of packing and repacking soil columns takes some time and suffers from the difficulty of repacking the soil to the same bulk density at each new water content (those that have tried this with heavier textured soils know what i mean). my method involves packing a soil column a single time, pumping water into the bottom of the column and taking simultaneous measurements of dielectric constant and mass of water pumped into the column. because TDR provides an average dielectric constant along the entire length of the probe, shart wetting fronts do not invalidate the measurement. we found, with 3 different textured soils, that the method is repeatable, fairly easy to set up, rapid (~12 hours for 20 cm probes), and capable of providing several hundred data points. scatter of data points around the fitted curve was very low, with RMSE values less than .67% volumetric water content in all cases. the paper has been approved for publication in the SSSAJ, and should be in print sometime in mid-1997. i can send a copy of the paper if you would like to see the paper earlier. i am also working on getting the manuscript on the WWW in Adobe Acrobat format. that should be ready by early next january. 2. one of my findings was that the slope and shape of my calibration curves for three soil textures and a 20 cm probe were very similar to topps 1980 curve (in fact, i found that they were statistically the same). his uncertainty of +/- 2% was for all his soils and other granular material, fitted simultaneously to a single calibration curve. steve evett's letter, posted a couple of days ago stated something that i also believe: that if you need true water content, with accuracy better than 2%, then you will probably have to calibrate your probes. however, if you're only looking at CHANGE in water content, using a commercially available probe with a 'normal' soil, then you might consider just using topps curve. 3. as i see it, resolution of the TDR system is dependent on the cable tester settings chosen by the user. that is, higher dist/div settings will lead to lower resolution because each data point on the trace corresponds to longer distances. thus, your resolution will decrease. the accuracy and precision would be derived from the calbration procedures. happy holidays to all. mike ------------------------------------------------------------------------- Michael H. Young, Ph.D., Assist. Res. Sci. Dept. Soil, Water and Env. Sci., Univ. of Arizona 429 Shantz Bldg. #38 Tucson, AZ 85721, USA phone: (520) 621-3236 fax: (520) 621-1647 email: youngmh@ag.arizona.edu