2.5. Soil moisture data
A TRIME-IPH TDR system (IMKO, Germany) was employed to sample soil moisture in each plot in the 0 depth range. In order to capture detailed temporal patterns of soil water use of jujube trees in the main root soil profiles, soil water content was also recorded using EC-5 sensors (Decagon Inc., USA) every 10 minutes in one plot for each of the four treatments. To install soil moisture sensors, a 320-cm-deep pit was dug and the soil moisture sensors were inserted horizontally at depths of 10, 20, 40, 60, 100, 160, 220, and under each of the four plots. Soil water data for the first month after sensor installation were excluded.
The soil moisture outputs (θE ) obtained via the EC-5 sensors were calibrated at the study site in order to improve accuracy. According to Li et al. (2018), the calibration equation was:
(1)
where denotes volumetric soil water content derived by multiplying gravimetric water content by dry soil bulk density.
In this study, daily soil water observations were used to analyze temporal patterns of water use by jujube trees. Daily soil water data clearly has less noise than the 10-minute and hourly data; however, using daily data can obscure the effect of gravity drainage following rainstorms when quantifying soil water use, because the relatively high infiltration capacity of loess soil means that gravity drainage is expected to last less than one day (Zhu et al., 1983). Daily soil moisture data were derived by arithmetically averaging soil moisture data at the 10-min resolution over a whole day.
Here, the 0-100 cm soil layer was defined as “shallow”, because this layer clearly exhibits higher temporal variations in soil moisture (Gao et al., 2014) and higher fine root density (Li et al., 2017) than the layers below 100 cm. Furthermore, this definition of a shallow layer can also be found in other regions (e.g., Broedel et al., 2017; Yang et al., 2017).