| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Soil Physics and the Environment Institute of Soil, Water, and the Environment Agricultural Research Organization Ministry of Agriculture, POB 6 Bet Dagan, Israel
vwnad{at}volcani.agri.gov.il
Chandler et al. (2004) tried to evaluate and improve the performance of the commercial soil moisture sensor CS-615 (Campbell Scientific Inc., Logan, UT). The authors should be complimented for their devoted efforts to provide growers with an accurate yet inexpensive tool for monitoring water content (WC), a critical parameter in agricultural production. Practically the authors propose a three-phase campaign:
My following comments intend to show that: (i) the calibration strategy offered and discussed by the authors may be problematic, (ii) the measurement frequency gap is unbridgeable, (iii) the selected experimental site is not a typical agricultural example, and (iv) it is worth checking if a simple and cheap electrical resistance measurement could not bring the same benefit for a fraction of the price.
The authors assume that "TDR is a reliable, most widely accepted electrical technique for measuring soil WC"; it is desirable to identify a substitute for the TDR technique due to its high price; and Campbell Scientific's Water Content Reflectometer (WCR, CS-615) is a suitable candidate because, similarly to the TDR it measures soil dielectric properties, differing only in the measurement frequency (15–45 MHz, compared with the 1000–1450 MHz of TDR).
Contrary to the superlatives used for describing TDR, the characteristics of the WCR presented in the article include (verbatim): "WCR-determined volumetric WC (VWC) is more sensitive to soil type, because the effects of EC are strongly temperature dependent ... WCR data are also temperature (T) dependent for high EC soils." For three of the four soils tested, the authors report, "there were substantial deviations from the factory calibration and substantial T effects." The authors also stated, "at present there are insufficient data to know, a priori, how much the calibration for a given WCR application might deviate from the factory calibration, however it appears that the calibration will vary with soil properties, with deviation tending to increase with clay content, and that it may vary with each individual sensor, and in addition, it is difficult to account for individual sensor variability, which may vary with site characteristics." Sensors' production seems also to have contributed to the error, as was indicated by the trends between individual sensor pairs (TDR-WCR), and the error in the overall relationship is primarily due to variability in the offset constant among the individual WCR sensors whose response may vary with production season or field installation or both.
The site-specific regressions show a difference in accuracy of WCR between the two sites, attributed by the authors to the difference in clay content. However, the 8% difference in clay content between the soils in both sites would not have made this difference with a true TDR. The authors note, "Much of the data for both sites fell near or outside the 3% error bounds. Assuming this range of error, the total measurement error of 6% WC is approximately 1/4 of the annual range in volumetric WC for these soils." Their Fig. 2, reporting WCTDR – WCWCR relations, conveniently demonstrates several types of disparity: partly or fully biased, systematically shifted, and a wide range of scatter, all relative to factory calibration.
To improve the correspondence between WCR and TDR the authors have applied different forms of averaging, apparently without considering whether the dual TDR-WCR installation can change the inherent properties of the WCR by calibration manipulations.
More specifically, it is given that the CS-615 is sensitive to T, EC, and the soil WC, which are continuously changing during the season, and that undistinguishable interactions exist between them. Thus the user is left with a problem of three unknowns and only one equation.
Experimental results show that the WCR response is predominantly determined by its sensitivity to the resistance of the medium it is imbedded in, including indirectly the complicated, nonlinear, multi-interdependence on T, texture, and salinity. Field measurements obtained by the growers may yield doubtful results, for only the texture can be expected to remain constant, whereas soluble salts and T may vary considerably during periods when TDR measurements are not taken. Consequently, if one adds up the individual sensor production variability and the random offset, or the unavoidable sensitivity to field parameters other than water content, it must be concluded that the CS-615 is no better than two bare metallic rods of a rigid configuration.
Based on the above comments, I would appreciate the authors addressing the following:
The experimental WC values obtained by the authors do not agree with the above hypothesis, and this may stem from their use of experimental conditions that are not representative of an agricultural field. The experimental site chosen by the authors is characterized by biasing conditions: a single sensor installed in a single location, a narrow range of water salinity, the absence of plants, and a rain fed field. The reader should note that such conditions do not reflect realistic agricultural situations where roots preferentially remove water but not salts, thus enhancing the EC effect.
As to the latter issue, it is surprising that temperature measurements were not reported by the authors. Moreover, if the authors suspect that a significantly different background EC was contributed by 8% clay content (difference between the two experimental sites), why did they not monitor the soil EC profile?
My comments cannot be complete without some mention of the article's humoristic aspect. The authors have concluded from their WCR data that the soil and inter-sensor variability can be effectively field-calibrated with TDR, resulting in WC measurements that approximate the accuracy of the TDR. They further conclude, "The ultimate limitation of the method is tied to the limitations of the TDR."
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
