Rapid percolation of water through soil macropores affects reading and calibration of large encapsulated TDR sensors

Abstract

Doležal F., Matula S., Moreira Barradas J.M. (2015): Rapid percolation of water through soil macropores affects reading and calibration of large encapsulated TDR sensors. Soil & Water Res., 10: 155-163.The electromagnetic soil water content sensors are invaluable tools because of their selective sensitivity to water, versatility, ease of automation and large resolution. A common drawback of most their types is their preferential sensitivity to water near to their surfaces. The ways in which the drawback manifests itself were explored for the case of large Time-Domain Reflectometry (TDR) sensors Aqua-Tel-TDR (Automata, Inc., now McCrometer CONNECT). Their field performance was investigated and compared with the results of field and laboratory calibration. The field soil was loamy Chernozem on a carbonate-rich loess substrate, while the laboratory calibration was done in fine quartz sand. In the field, the sensors were installed horizontally into pre-bored holes after being wrapped in slurry of native soil or fine earth. Large sensor-to-sensor variability of readings was observed. It was partially removed by field calibration. The occurrence of percolation events could be easily recognised, because they made the TDR readings suddenly rising and sometimes considerably exceeding the saturated water content. After the events, the TDR readings fell, usually equally suddenly, remaining afterwards at the levels somewhat higher than those before the event. These phenomena can be explained by the preferential flow of water in natural and artificial soil macropores around the sensors. It is hypothesised that the percolating water which enters the gaps and other voids around the sensors accumulates there for short time, being hindered by the sensors themselves. This water also has a enlarged opportunity to get absorbed by the adjacent soil matrix. The variance of TDR readings obtained during the field calibration does not differ significantly from the variance of the corresponding gravimetric sampling data. This suggests that the slope of the field calibration equation is close to unity, in contrast to the laboratory calibration in quartz sand. This difference in slopes can be explained by the presence or absence, respectively, of gaps around the sensors. A typical percolation event and dry period records are presented and analysed. Sensors of this type can be used for qualitative detection of preferential flow and perhaps also for its quantification. The readings outside the percolation events indicate that the sensor environment imitates the native soil reasonably well and that the field-calibrated sensors can provide us with quantitative information about the actual soil water content.

Preferential percolation quantified by large water content sensors with artifactual macroporous envelopes

“…In our case, the field calibration (Doležal et al ., ; ) consisted in relating the sensor readings in cm 3 cm −3 to the soil–water contents obtained gravimetrically at a distance of the order of 1 m from the sensors. While the laboratory calibration of another AQUA‐TEL‐TDR sensor in quartz sand with volumetric water contents from zero to 0.41 cm 3 cm −3 (Doležal et al ., ) resulted in a virtually linear calibration equation with a slope (sampling vs TDR) 0.654 and no anomalies ( R 2 = 0.992), the field calibration in our loamy soil showed virtually no correlation between the sampled and TDR values related to individual sensors (with R 2 less than 0.042 and the regression coefficients between −0.267 and −0.002), because of the small‐scale heterogeneity of the native soil–water content, captured by gravimetry but averaged by the large TDR sensors.…”

“…A more detailed description of the experimental area, field measurements and sensor calibration procedures was provided by Doležal et al . (,b; ).…”

“…We build on the results obtained by Doležal et al . (,b; ) and on their conceptual model of a thin annular macroporous zone around the sensors, in which the percolating water accumulates for a while. In particular, we want to find out whether or not the high and short‐lasting peaks obtained by large horizontal TDR water content sensors and the processes that these data represent can be feasibly simulated by a suitable soil–water model; the model parameters needed for such simulation can be reasonably well estimated, and, once they have been estimated, the simulations can predict rapid percolation fluxes at particular depths from the known rain intensities. …”

“…Doležal et al . (,b; ) conclude that the anomalous and short‐lasting water content peaks are indicators of the rapid downward percolation, which is very probably of preferential (i.e. non‐equilibrium) nature.…”

“…Doležal et al . (,b; ) measured soil–water content by large time‐domain reflectometry (TDR) sensors that had been installed horizontally in a loamy Udic Haplustoll. They found that the TDR readings become very high for some time during and after intensive rainfall and snowmelt events.…”