2.7 Dendrometers
Linear variable displacement transducers were used as high-precision point dendrometers (ZN11, Zweifel Consulting, Hombrechtikon, Switzerland). Technical information about the sensors and their mounting to tree stems can be found at Zweifel et al. (2014). Pairs of dendrometer units were installed at 60 cm above ground (termed ‘base’) and at a position where stem diameter was 3 cm (termed ‘top’), resulting in a total of 24 mounted units. Installation height of dendrometers at top varied among trees (Table 1), but distance between base and top was always more than 2.5 m. Pistons of the dendrometers were positioned directly on bark and sapwood. For the latter, we cut a small bark window (1 cm2) using a chisel. Bark windows were flushed with deionised water after carefully scraping cambial cells from the outer sapwood. Once cleaned and padded dry, the tip of a dendrometer plunger was positioned directly onto sapwood and bark windows were sealed using silicone grease to prevent drying of exposed stem tissues.
Measurements recorded on sapwood were used to calculate the amplitude of radial movement in xylem (D x). SubtractingD x from measurements collected on bark isolates the bark signal representing radial movement generated by reversible (cambium and phloem) and irreversible (formation of wood) radial movement. In a second step, we separated radial movement related to the two co-occurring processes. First, we calculated the difference between maximal and minimal dimensions during a single diurnal cycle to represent the reversible, bark-driven proportion of the total movement (D b) only. Next, we assessed the difference between two consecutive minimum expansions measured on bark to determine irreversible cambial growth (D Gempir). These are the simplest mathematical procedures when analysing data from simultaneous measurements on sapwood and bark using point dendrometers (Mencucciniet al. 2017). We note that the termD Gempir is a composite measure that contains hydraulic and osmotic signals that can mask true irreversible cambial growth. See section ‘Stem Hydraulic Model’below for further details.
To assess the effect of girdling on dimensional changes of live bark and xylem, we zeroed dendrometer readings against the measurement recorded at midnight of 5 May 2014 (pre-girdling) and 19 May (post-girdling). Time-series data (base and top positions of all six trees) from 6-10 May was used as baseline data, measurements form 20-24 May were used to assess how girdling affected D x,D b andD Gempir. The prefix ‘d’ will be used to signify ‘change in’ (e.g., dD x is the dimensional change in radial direction in xylem).