Comments on "J. Vera et. al., Soil water balance trial involving capacitance and neutron probe measurements"

Authors: Evett, Steven - Steve; Schwartz, Robert

Submitted to: Agricultural Water Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/1/2010
Publication Date: 4/16/2010
Citation: Evett, S.R., Schwartz, R.C. 2010. Comments on "J. Vera et. al., Soil water balance trial involving capacitance and neutron probe measurements". Agricultural Water Management. 96:905-911, Agricultural Water Management. 97:182-184.

Interpretive Summary: In this paper, scientists of the USDA-ARS Soil and Water Management Research Unit, Bushland, Texas, refute incorrect conclusions made in a paper by other authors. The incorrect conclusions indicated that there were no important differences between two technologies for sensing soil water content (and thus crop water use) and that the choice between these technologies could be made on the basis of price and convenience alone. Such water sensing technologies are key in efforts to improve water management, increase the efficiency with which water is used in irrigated agricultural systems, and improve the management of soil water and cropping systems in dryland agriculture for increased yields. Choice of the wrong soil water sensing technology can lead to failures in irrigation and crop management that are very costly to producers. Using physical and statistical analysis, it was shown that indeed there were important differences between the sensors, and that one sensor type (the neutron moisture meter) was more accurate and was capable of accurately determining crop water use. It was shown that the other sensor, which was based on electronic sensing, was made inaccurate by soil salinity caused by irrigation. This paper removes incorrect conclusions from the scientific literature and makes it less likely that the wrong sensor type will be chosen for important research and agricultural production problems.

Technical Abstract: Vera et al. (2009) compared estimates of soil profile water content (mm) to a depth of 0.8 m made with the neutron moisture meter (NMM) and a multi-depth capacitance probe (MDCP), using measurements replicated in four drainage lysimeters (5 m x 5 m x 1.5-m deep). The NMM estimates of water content were reasonable and less than the saturated water content of the soil. The MDCP estimates were larger than the saturated water content value and therefore were physically unrealistic and erroneous. Clear effects of soil bulk electrical conductivity (BEC) on the MDCP estimates of soil water content were visible in the graphs published by Vera et al. (2009), but these effects and the physically unrealistic soil water content estimates were not acknowledged by those authors who concluded that there were no important differences between the NMM and the MDCP. In the present manuscript, these errors are pointed out using physically based analysis and modeling of the effects of BEC on estimates of soil electrical permittivity and thus on estimates of water content by electromagnetic sensors operating in the same frequency range as the MDCP. The conclusions of Vera et al. (2009) are refuted using these analyses and by examining the relative statistical power of comparisons between lysimeter-derived changes in soil water storage (delta-S) and NMM-derived estimates of delta-S on the one hand, and between lysimeter-derived delta-S and MDCP-derived delta-S on the other hand. Since the standard deviations of NMM estimates of water content were much smaller than those of MDCP estimates of water content, the statistical power of the comparison between lysimeter-derived delta-S and NMM-derived delta-S is great, meaning that the lack of statistically significant difference between these estimates of delta-S leads to the conclusion that the NMM-derived delta-S values are the same as the lysimeter-derived values. That is, the NMM is a good tool for determining water contents and changes in profile water content over time, and thus is a good tool for soil water balance estimates of crop water use. The large standard deviations of MDCP estimates of water content lead to considerably less statistical power for the comparison of MDCP-derived delta-S and lysimeter-derived delta-S. Thus, the lack of a statistically significant difference in this case does not necessarily indicate that the MDCP is a good tool for determining delta-S.