Changes in ultrasound velocity and attenuation indicate freezing of xylem sap

Agricultural and Forest Meteorology 185 (2014) 20–25

Contents lists available at ScienceDirect

Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet

Technical Note

Changes in ultrasound velocity and attenuation indicate freezing of xylem sap Charrier Guillaume a,∗ , Charra-Vaskou Katline b,c , Legros Benoit d , Améglio Thierry b,c , Mayr Stefan a a

Department of Botany, University of Innsbruck, Sternwartestr. 15, A-6020 Innsbruck, Austria INRA, UMR547 PIAF, F-63100 Clermont-Ferrand, France c Clermont Université, UMR547 PIAF, F-63 Aubiere, France d MISTRAS Group, SA Division Euro Physical Acoustics, 27 Rue Magellan – ZAC des Portes de Sucy, F 94370 Sucy en Brie, France b

a r t i c l e

i n f o

Article history: Received 13 June 2013 Received in revised form 14 October 2013 Accepted 20 October 2013 Keywords: Freezing Ice Sound velocity Ultrasound propagation Wave attenuation Xylem

a b s t r a c t Freezing is a limiting factor for plant life, as it can cause damage of living tissues and embolism formation in conduits. Ice formation in plant tissues is usually detected by exotherm analysis. In this study, a new method based on changes in ultrasonic properties of wood was used to monitor xylem freezing. Ultrasound propagation velocities and attenuation were measured with an ultrasonic emission analysis system in branches of three conifer and three angiosperm tree species by signal induction via auto sensor test (AST) or lead break (LB: Hsu–Nielsen source). In all species under study, ultrasound velocity was 1.2–3.6-times higher in frozen xylem (−10 ◦ C) compared to samples at 10 ◦ C. In Picea abies, velocities of AST signals increased from 2193 to 3085 m s−1 and in Fagus sylvatica from 2369 to 3009 m s−1 , while signal attenuation decreased in both species. The crystalline structure of ice with slower molecular movements and strong hydrogen bonding caused the faster propagation and reduced attenuation of acoustic waves after xylem freezing. Xylem anatomy also influenced acoustic properties as demonstrated by inter-species differences in temperature responses. The analysis of ultrasonic properties provides a new method for the detection of ice in the xylem of trees and may be used to monitor stress intensities and estimate physiological effects in situ. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Freezing is a major limiting factor for plant life in several environments, such as temperate and boreal biomes or alpine areas. In the xylem of woody plants, freezing can lead to damage by different mechanisms. Due to the volume increase of freezing sap, mechanical constraints can exceed cell wall rigidity and provoke frost cracks (Ishida, 1963; Cinotti, 1991). Ice formation also causes dehydration and osmotic stress affecting the cytoplasm and membranes of living cells (Steponkus, 1981; Ruelland et al., 2009). Extreme temperatures below −30 ◦ C can provoke intracellular ice formation even in deep supercooling species (Fujikawa and Kuroda, 2000), which is always lethal for cells (Wolfe and Bryant, 2001). Plants also suffer from freeze–thaw cycles, when embolism is induced in

Abbreviations: AST, automated sensor test; LB, lead break, refers to Hsu–Nielsen simulation test. ∗ Corresponding author. Tel.: +43 512 507 51029. E-mail addresses: [email protected] (C. Guillaume), [email protected] (C.-V. Katline), [email protected] (L. Benoit), [email protected] (A. Thierry), [email protected] (M. Stefan). 0168-1923/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.agrformet.2013.10.009

xylem conduits (Sperry and Sullivan, 1992; Tyree et al., 1994; Hacke et al., 2001; Mayr et al., 2007). Within a plant, freezing susceptibility differs between tissues so that dynamics of ice formation play an important role in development and extent of damages. Similarly, the pattern of ice in the xylem determines hydraulic blockages. Monitoring freezing and temporal and spatial dynamics of ice formation in plants are thus a prerequisite to evaluate stress intensities and their physiological effects in situ. Freezing in plant tissues can be monitored by high resolution temperature measurements, which enable the detection of exotherms. When liquid water turns into ice, the molecular structure changes a crystalline formation. This causes a release of latent heat energy (L = 334 J g−1 ), which can be recorded as an exotherm (Muldrew et al., 2004). When apoplastic water freezes, a first high temperature exotherm is observed. When the temperature decreases further, a second exotherm corresponding to intracellular ice nucleation appears. The cell sap freezes at lower temperatures because the higher solute concentration causes a deeper freezing point (Kasuga et al., 2007). Numerous studies based on exotherm analysis enabled insights into the freezing process and frost resistance (e.g. Burke et al., 1976; Kuroda et al., 1999), but in situ, the detection of exotherms is difficult due to rapid dispersion of released heat, especially under turbulent atmospheric

C. Guillaume et al. / Agricultural and Forest Meteorology 185 (2014) 20–25

conditions. The first attempts were performed by Ashworth et al. (1985) on peach trees: differential thermal analysis enabled the detection of temperature differences at high resolution (Fujikawa et al., 1994; Pramsohler et al., 2012). Exotherms can be visualised also via infra-red thermography (Wisniewski et al., 1997). Up to now, this technique has only been used in laboratory experiments (e.g. Hacker and Neuner, 2008) and it allows temperature analysis only at the sample surface. In a few studies, time domain reflectrometry was used to detect ice fractions in stems (Sparks et al., 2000). Sparks et al. (2001) found portions of liquid water in Pinus contorta stems even in deep winter. Effect of freeze–thaw stress can also be monitored via acoustic emission analysis. In several studies (Kikuta and Richter, 2003; Mayr et al., 2007; Mayr and Zublasing, 2010; Mayr and Sperry, 2010) acoustic emissions were registered during freezing and found to be correlated with the loss of hydraulic conductivity in the xylem. Up to now, acoustic emission analysis on plants predominantly focused on the number and dynamics of signals induced by drought (Tyree and Dixon, 1983; Salleo and Lo Gullo, 1986) or freeze–thaw events (Raschi et al., 1989). The signal quality was hardly considered (Mayr and Rosner, 2011) although it may give important additional information. Sound waves are mechanical vibrations creating periodic disturbances as a succession of compression/rarefaction. The ability of a solid material to allow sound propagation depends on its elasticity and density (Lampriere, 2001):



Cl =

1− × 2(1 + )(1 − 2)



E 

(1)

where Cl is the propagation velocity of sound waves in longitudinal direction,  is the Poisson’s ratio,  is the density of the material and E is the Young’s modulus. In a heterogeneous material like wood, with solid, liquid and even gases, sound propagation is more complex. It depends on density, elasticity, adiabatic compressibility coefficient (for liquids) and, because it is an orthotropic material, direction of the propagation (Bucur, 2006). Inter-species variability in sound propagation velocities is directly related to Young’s modulus and density (Kretschmann, 2010). Temperature also influences sound propagation velocity because of its specific effect on the elasticity and density of the medium. In liquids and gases, sound velocity is proportional to temperature, whereas in solids, increasing temperatures induce a decrease in Young’s modulus and, in consequence, in velocity. During the transition of water to ice, the density of the medium decreases and Young’s modulus increases. In consequence, sound propagation velocity increases from ca. 1480 m s−1 in water (Bilaniuk and Wong, 1993) to ca. 3900 m s−1 , in ice (Smith and Kishoni, 1986). We hypothesised that the ultrasound propagation velocity and other qualitative parameters (e.g. amplitude, attenuation) of ultrasonic signals might be used as an indicator for the phase of water in the xylem. In this study, we tested the effect of freezing on ultrasound velocity, amplitude and attenuation in stems of several tree species. We induced artificial ultrasonic signals and expected that (i) ultrasound propagation velocities in frozen samples were higher and the attenuation lower than in unfrozen samples and that (ii) differences among species were small as the sound propagation predominantly occurs via the water column. 2. Material and methods 2.1. Plants Branches were sampled from mature trees growing near the Department of Botany in Innsbruck, Austria. Measurements were

21

Fig. 1. Scheme of sample dimensions, sensor and lead break positions (pencil in grey): total length L = 30 cm; diameter d ca. 1 cm; distance to lead break L1 = 2 cm (conifers) or 4 cm (angiosperms); distance between sensors L2 = 4 cm (conifers) and 8 cm (angiosperms); minimal distance to the end L3 ≥ 5 cm. Bark is in grey and wood in white. In experiment 1, all four sensors were present (radial velocity was calculated between 1 and 3 or 2 and 4; axial between 1 and 2 or 3 and 4). In experiment 2 and 4 only sensor 1 and 2 were present. In experiment 3, only one sensor was present and different distances to LB source were used (L1 : 1–25 cm).

performed on three angiosperm (Acer pseudoplatanus L. (n = 3), Fagus sylvatica L. (n = 8) and Juglans regia L. (n = 3)) and three conifer species (Larix decidua Mill. (n = 3), Picea abies (L.) Karst. (n = 8) and Pinus sylvestris L. (n = 3)). Branches were approximately 1–2 cm in basal diameter and 60 cm in length. From branches, 30 cm long segments of the main stem were used. The mean vessel length of all species was shorter than 30 cm and thus shorter than the sample length. Side branches (if present) were removed as the space in the temperature chamber was limited. Segments were cut out of the stem under water and flushed with water (a.d.) at 0.16 MPa for 10 min to release tension and remove xylem embolism (enclosed air might affect the freezing process by local insulation and slow down sound propagation). Samples were wrapped into parafilm (Alcan, Montreal, Canada) to prevent dehydration. 2.2. Freeze–thaw cycles Freeze–thaw cycles were performed in a temperature test chamber (MK53, Binder GmbH, Tuttlingen, Germany). Within a frost cycle, temperature decreased from +10 to −10 ◦ C at 5 K h−1 and stayed one hour at minimal temperature before thawing at 5 K h−1 . Xylem and air temperature were monitored using copperconstantan thermocouples (one per sample, on one end, 2 cm from an acoustic sensor) connected to a datalogger (CR10X, Campbell Scientific Ltd., England). To measure temperature effects in ice, some experiments were performed down to −20 ◦ C at identical freezing and thawing rates in Fagus and Picea. 2.3. Ultrasound propagation velocity Ultrasonic measurements were performed with a PCI-2-based system (PAC125, 18-bitA/D, 3 kHz–3 MHz) and 150 kHz resonance sensors (R15) connected to a preamplifier set to 40 dB (all components: Physical Acoustics Deutschland, Wolfegg, Germany). The threshold was set to 45 dBEA (0 dBEA = 1 ␮V; e.g. Mayr et al., 2007; Mayr and Rosner, 2011). Registration and analysis of ultrasonic events were done with AEwin software (Mistras Holdings Corp., Princeton, USA). About 1 cm2 of bark (and parafilm) were removed from the samples and the xylem was covered with silicone grease (to ensure acoustic coupling and prevent dehydration) before attaching the sensors with clamps. Sensors were placed at a distance of 8 cm on angiosperm stems and of 4 cm on conifers because of higher attenuation in coniferous species (Fig. 1). Ultrasound velocity was calculated from the distance between sensors divided by the time difference of corresponding hits. Temporal resolution of the acquisition system is ca. 0.25 ␮s, which gives high accuracy for velocity measurements. Ultrasonic signals were

C. Guillaume et al. / Agricultural and Forest Meteorology 185 (2014) 20–25

15

A

3000

10 2500

5

2000

0 -5

1500

-15

Picea abies

15

B

3000

10 2500

5 0

2000

-5

1500

-10

Axial Velocity Radial Velocity

1000

0

(1) Velocity was measured by AST every 15 min during two freeze–thaw cycles between +10 and −10 ◦ C in Picea and Fagus on 1 replicate per species. Four sensors were placed on the sample, two on one side (compression wood side, 4 cm distance in Picea, 8 cm in Fagus) and two on the opposite side for radial (between sensors 1 and 3 as well as 2 and 4) and axial velocity measurements (between sensors 1 and 2 as well as 3 and 4). (2) Velocity and signal amplitude were measured by AST and LB during a temperature decrease from +10 to −20 ◦ C with one hour step every 5 ◦ C in Picea and Fagus on 7 replicates per species. Two sensors were positioned at the upper side (opposite wood) of the samples. (3) Amplitude of the signal depending on the distance was measured by LB. One sensor was positioned at the upper side (opposite wood) of the samples, 5 cm from one end. LB tests were made every centimetre (towards the sample centre) from a sensor at +10, −10 and −20 ◦ C in Picea and Fagus on 5 replicates per species. (4) Velocity and signal amplitude were measured by AST and LB in six species (from 7 replicates for Picea and Fagus and 3 replicates for Acer, Juglans, Larix and Pinus) at +10 and −10 ◦ C. Two sensors were positioned at the upper side of the samples.

-10

Xylem Temperature Air Temperature

1000

Temperature (°C)

produced either by Automated Sensor Test (AST) of the ultrasonic analysis system or by Hsu–Nielsen simulation test (NF EN1330-9) produced by a black lead break (LB) on the bark surface. During AST, one sensor emitted an acoustic signal, directly on the wood surface, which was received by another sensor. Propagation times calculated with this method are highly reproducible, because the acoustic signal has a constant frequency (peak at 150 kHz). On each sample, we performed 20 pulses (5 ␮s duration and 100 ms between pulses). LB is a standard method (Hsu–Nielsen lead break) for creating ultrasonic signals with relatively constant amplitude, which is used to test sensor coupling or attenuation properties of a material (Kalyanasundaram et al., 2007). This method produces signals similar to intrinsic emissions, especially regarding frequencies. The distance between the transducer and the Hsu–Nielsen lead break source was 2 cm in conifers and 4 cm in angiosperms. For each sample, the velocity was calculated as the mean of 5 different tests. The following experiments were performed (see also Fig. 1):

Ultrasound propagation velocity (m.s -1)

22

Fagus sylvatica 4

8

-15

12

Time (h) Fig. 2. Longitudinal and radial velocities of ultrasound propagation in Picea abies (a) and Fagus sylvatica (b) during two freeze–thaw cycles. Automated sensor test (AST) was used to measure sound velocities in axial (squares) and radial (circles) directions with four sensors attached to stems (mean ± SE; n = 4). Lines show temperatures in the air (black) and xylem (grey), arrows indicate freezing exotherm.

velocity only between −20 and +10 ◦ C. With LB, a difference was already observed between 0 and −5 ◦ C. Velocities calculated with AST and LB were not significantly different from each other, except for +5 ◦ C. No difference was observed in amplitudes of AST and LB induced signals, and the temperature dependent pattern was similar to velocity patterns. In contrast to Picea, velocities and amplitudes were lower in AST than in LB signals in Fagus and changes between 0 and −5 ◦ C were more pronounced. Signal amplitudes decreased with distance from the sensor (Fig. 4), with lower attenuation at lower temperatures. In Picea, initial amplitude of the signal decreased by 30 dB (d−30 ) at a distance of 3.0 cm between the ultrasonic source and the transducer and signals could be detected to a maximum distance of 15 cm. In frozen samples, signals were detected at up to 20 cm distance. Decreases of 30 dB from initial amplitude were reached at a distance of 5.0 cm (−10 ◦ C) and of 10.1 cm (−20 ◦ C). In Fagus, decreases of 30 dB from

2.4. Statistical analysis

During temperature decrease, velocities of AST induced signals remained similar until the onset of freezing (indicated by an exotherm monitored with thermocouples; arrow in Fig. 2). After complete freezing and reach of the final temperature (−10 ◦ C in experiment 1), velocities were 20–25% higher than in unfrozen samples and remained constant until thawing. On thaw, velocities decreased to previous values. An identical pattern was observed during a second freeze thaw cycle. In radial direction, velocities were similar in Picea and Fagus (1469 ± 5 m s−1 and 1478 ± 2 m s−1 ). Significant differences and overall higher velocities were observed in axial direction (2150 ± 117 m s−1 in Picea and 1950 ± 63 m s−1 in Fagus). Axial velocity measurements obtained with the AST method were compared with the LB method at different temperatures (Fig. 3). In Picea, AST signals revealed a significant difference in

Velocity (m.s -1)

3. Results

Picea abies 3500

Amplitude (dB)

After testing for Gaussian distribution of velocity and amplitude values, ANOVA analyses and post hoc Student’s t-tests were performed to segregate statistically significant groups (˛ = 0.05) using R software (R Development Core Team, 2005).

a

Fagus sylvatica A

a a

3000

ab a ab

2500 2000

ab

a

LB AST a

90

ab

b

* ab

b a

* b

B a

a

a

*

*

*

a

a

a

b

b

b

*

*

b

b

C a

ab ab

ab

bc c

70 LB AST

-20 -15 -10 -5

cd d

0

cd d

5

d

a

*

*

*

a

a

a

* b

b

*

c

a

*

d

10

b

D a

b

a

80

60

a

-20 -15 -10 -5

cd

*

b

b

0

5

d

* b

10

Temperature (°C) Fig. 3. Axial sound propagation velocities (A and B) and amplitudes (C and D) with Lead Break (LB: full symbol) and automated sensor test (AST: open symbol) in Picea abies and Fagus sylvatica at different temperatures (mean ± SE; n = 7). Different letters represent significant differences among temperatures (P < 0.05) for LB and AST (italic). Stars indicate significant difference between AST and LB for a temperature, respectively.

C. Guillaume et al. / Agricultural and Forest Meteorology 185 (2014) 20–25

Fig. 4. Attenuation of ultrasound signals induced by lead break (LB) along branches of Picea abies (a) and Fagus sylvatica (b) at different temperatures: unfrozen (+10 ◦ C) or frozen (−10 and −20 ◦ C) (mean ± SE; n = 5). Lines represent the best fit with exponential function and arrows represent the distance where initial amplitude of the signal decreased by 30 dB (d−30 ).

initial amplitude were observed at a distance of 10.8 cm (+10 ◦ C), 17.7 cm (−10 ◦ C) and 21.6 cm (−20 ◦ C). A significant increase in ultrasound propagation velocity upon freezing was also observed in other angiosperms (Juglans regia and Acer pseudoplatanus) and conifers (Pinus sylvestris and Larix decidua) under study (Table 1). The ratio of velocities in frozen vs. unfrozen samples ranged from 1.2 up to 3.6. Velocities measured with AST were overall lower than with LB, except in unfrozen conifers. Significant differences between AST and LB were observed in unfrozen Pinus and Fagus and in frozen Pinus, Fagus, Juglans and Acer samples (Table 1). 4. Discussion An increase in ultrasound velocity was observed in all species when samples started to freeze (Table. 1). This change in velocity obviously corresponded to the liquid to solid phase transition of the xylem sap. While the liquid state is characterised by hydrogen bonds of short lifespan (ca. 1 ps) and few bonds per molecule (only about 25% of molecules are connected by four hydrogen bonds and 25% are unbound at 0 ◦ C; Teixeira and Luzar, 1999), steady bonds with four neighbour molecules build a crystalline structure in ice. This rigid structure supports sound propagation from one molecule to the other and increases propagation velocities (Bucur, 2006). Between 0 and −10 ◦ C, water is under crystallisation and the velocity increase depends on the amount of ice (Gülseren and Coupland, 2007). This explains intermediate values observed between the unfrozen and the frozen state (Fig. 3). The velocity of an ultrasonic wave propagating longitudinally through a material mostly depends on Young’s modulus and the density of the material (Lampriere, 2001). Young’s modulus exhibit a significant increase (from 13% to 300%) in frozen branches compared to thawed, especially in conifers (Umbanhowar et al., 2008). An increase in Young’s Modulus will result in an increase of the ultrasonic wave velocity. Like in aluminium or steel (Weaver and Lobkiss, 2001), decreasing temperatures in wood cause an increase in Young’ Modulus (i.e. increasing stiffness) and, in consequence, in acoustic wave velocity (James et al., 1982; van Dyk and Rice, 2005). Also the Young’s Modulus of ice increases when the temperature decreases (Lindgren, 1970; Kong and Campbell, 1987), which corresponds to changes in velocities with decreasing temperature in frozen samples (Fig. 3). The density of ice is lower than of liquid water (917 and 1000 kg m−3 , respectively), which

23

also contributes to the increase in velocity within the water column. In wood, moisture content has also been shown to play a role for acoustic properties (Sakai et al., 1990), such as sound propagation velocity (van Dyk and Rice, 2005; Auge, 1990; Bucur, 2006) or attenuation (Saadat-Nia et al., 2011) due to its effect on Young’s modulus and density (Kretschmann, 2010). In our study, samples were flushed with water before temperature treatments. Thus, saturation was ensured in all samples and an influence of moisture content could be excluded. We observed similar shifts to higher velocities on freezing regardless of the method with which signals were induced (Fig. 2 and Table 1). However, differences in the extent of velocity responses across methods were found in Fagus, Pinus and frozen Acer or Juglans. We suggest that differences are due to qualities of AST and LB signals. With LB, broader and higher frequencies are produced, which propagate faster but are also more attenuated. Attenuation increases with frequency, especially in composite materials (Kalyanasundaram et al., 2007). Moreover, AST signals, directly generated on the wood, could propagate via different ways than LB signals produced at the bark surface. AST might be useful as it might be automated. We also observed more pronounced interspecific differences in unfrozen than in frozen samples (Table 1) and large differences especially across conifers. Wood of conifers is composed of small tracheid elements (10–20 ␮m width), forming successive layers of cell walls and lumina, which may filter and attenuate acoustic waves more than angiosperms (from 20 to 100 ␮m). Wood anatomy and mechanical properties (cell wall thickness, density, Young’s modulus, presence and size of vessels and fibre length) thus has a significant effect on velocity (Huang et al., 2003; Bucur, 2006; Kretschmann, 2010). In general, wood densities and Young’s modulus are lower in conifers than in angiosperms (Kretschmann, 2010), but these parameters were not determined for fully hydrated branches under study. When frozen, Young’s modulus in conifers exhibit higher increase in mean and variance than in angiosperms (Umbanhowar et al., 2008). Hearmon (1948) reported a 30 times difference in Young’s modulus in axial direction across species. Young’s modulus thereby is generally lower (5% to 50% of longitudinal direction’s; Hearmon, 1948) explaining also the lower velocity in radial direction observed in Fig. 2. As for velocity, attenuation of ultrasonic signals substantially decreased with freezing but also showed some changes with temperature in frozen and unfrozen samples (Figs. 3 and 4). Crystalline structure and slower molecular movement enables a more efficient acoustic transmission of signals from one molecule to the other similar to the described influence on sound velocity (see above and Bucur, 2006). Acoustic signals thus can spread over longer distances (Fig. 4). Xylem anatomy also influences attenuation properties and is the reason for differences across species: d−30 was two to three times lower in Picea than in Fagus due to the composition of smaller tracheids (see above). Monitoring sound propagation velocities and attenuation enables new possibilities to study freezing and thawing in situ. This method might be used under environmental conditions, which do not allow exotherm detection (e.g. windy situations). In contrast to exotherm techniques, which allow detection of freezing only at the moment of ice crystallisation, this method enables detection of the freezing process and of ice after freezing. Furthermore, this method does not require high temporal resolution, while exotherm detection is only possible when temperature data acquisition occurs in short intervals (seconds). It may also be used for long-term monitoring (whereby automated AST measurements would be preferable) and enable detection of ice formation and melting as well as intrinsic acoustic emissions caused by embolism formation in parallel. For the detection of the onset of freezing, sensors should preferentially be placed on distal branches, which

24

C. Guillaume et al. / Agricultural and Forest Meteorology 185 (2014) 20–25

Table 1 Comparison of the ultrasound propagation velocity (m s−1 ) in three conifers and three angiosperms in unfrozen (+10 ◦ C) or frozen (−10 ◦ C) state using lead break (LB) or auto sensor test (AST): mean ± SE; n = 7 (Picea and Fagus) or n = 3 (other species). Increasing ratio of frozen and unfrozen samples and P-values are given. Species

Temperature

LB

Picea abies

+10 ◦ C −10 ◦ C +10 ◦ C −10 ◦ C +10 ◦ C −10 ◦ C +10 ◦ C −10 ◦ C +10 ◦ C −10 ◦ C +10 ◦ C −10 ◦ C

2193 3085 714 2553 1067 2400 2369 3009 2037 3088 2231 2969

Pinus sylvestris Larix decidua Fagus sylvatica Acer pseudoplatanus Juglans regia

± ± ± ± ± ± ± ± ± ± ± ±

151.7 106.8 62.5 88.6 267.5 24.3 65.1 91.3 195.1 269.8 222.5 126.8

Ratio

P-value

AST

1.41

<0.001

3.58

<0.001

2.25

0.008

1.27

<0.001

1.52

0.013

1.36

0.045

2381 2829 890 1893 1506 2476 1909 2499 1735 2158 1268 2042

are exposed to lower temperatures and often freeze before main branches. Use of several sensors in parallel will also enable insights into spatial and temporal dynamics of the freezing and thawing process in situ.

5. Conclusions The change in ultrasound propagation velocity and attenuation on freezing observed in all species under study was significant. These patterns might be used to detect spatial and temporal dynamics of freezing in plant stems. Based on AST, an automated monitoring, even under field conditions, is possible. Interspecific differences in acoustic properties demonstrate that xylem anatomy plays a crucial role for observed acoustic patterns so that test measurements are recommended before acoustic analysis of xylem ice.

Acknowledgements This study was supported by the Austrian Science Fund (FWF), I826-B25 and the Agence Nationale de la Recherche in the Acoufreeze Project (2012–2014).

References Ashworth, E.N., Anderson, J.A., Davis, G.A., Lightner, G.W., 1985. Ice formation in Prunus persica under field conditions. J. Am. Soc. Hort. Sci. 110, 322–324. Auge, F., 1990. Influence de l humidité du bois et de sa température sur la propagation des ultrasons. Nancy I University, Master Thesis. Bilaniuk, N., Wong, G.S.K., 1993. Speed of sound in pure water as a function of temperature. J. Acoust. Soc. Am. 93, 1609–1612. Bucur, V., 2006. Acoustics of Wood. CRC Press, New York. Burke, M.J., Gusta, L.V., Quamme, H.A., Weiser, C.J., Li, P.H., 1976. Freezing and injury in plants. Annu. Rev. Plant Phys. 27, 507–528. Cinotti, B., 1991. Investigation of the intrinsic properties of wood for explaining the occurrence of frost crack in Quercus petraea (Liebl) and Quercus robur (L) Ann. For. Sci. 48, 453–468. Fujikawa, S., Kuroda, K., 2000. Cryo-scanning electron microscopic study on freezing behavior of xylem ray parenchyma cells in hardwood species. Micron 31, 669–686. Fujikawa, S., Kuroda, K., Fukazawa, K., 1994. Ultrastructural study of deep supercooling of xylem ray parenchyma cells from Styrax obassia. Micron 25, 241–252. Gülseren, I., Coupland, J.N., 2007. Ultrasonic velocity measurements in frozen model food solutions. J. Food Eng. 79, 1071–1078. Hacke, U.G., Stiller, V., Sperry, J.S., Pittermann, J., McCulloh, K.A., 2001. Cavitation fatigue embolism and refilling cycles can weaken the cavitation resistance of xylem. Plant Physiol. 125, 779–786. Hacker, J., Neuner, G., 2008. Ice propagation in dehardened alpine plant species studied by infrared differential thermal analysis (IDTA). Art. Antarct. Alp. Res. 40, 660–670. Hearmon, R.F.S., 1948. The elasticity of wood and plywood. In: Forest Products Research Special Rep. 7. Department of Science and Industry Research, HMSO, London. Huang, C.L., Lindstrom, H., Nakada, R., Ralston, J., 2003. Cell wall structure and wood properties determined by acoustics – a selective review. Holz. Als. Roh. Werk. 61, 321–335.

± ± ± ± ± ± ± ± ± ± ± ±

123.2 175.5 31.9 80.3 65.8 24.5 43.7 54.4 72.9 18.1 304.5 149.5

Ratio

P-value

1.19

0.046

2.13

<0.001

1.64

<0.001

1.31

<0.001

1.24

<0.001

1.61

0.029

Ishida, S., 1963. On the development of frost cracks on Todomatsu trunks, Abies sachalinensis especially in relation to their wetwood. Res. Bull. Col. Exp. For. 22, 273–373, Hokkaido University. James, W.L., Boone, R.S., Galligan, W., 1982. Using speed of sound in wood to monitor drying in a kiln. For. Prod. J. 32, 27–34. Kalyanasundaram, P., Mukhopadhyay, C.K., Subba Rao, S.V., 2007. Practical Acoustic Emission National Certification Board Series Indian Society for Non-Destructive Testing – National Certification Board series. Alpha Science International, Oxford, UK. Kasuga, J., Arakawa, K., Fujikawa, S., 2007. High accumulation of soluble sugars in deep supercooling Japanese white birch xylem parenchyma cells. New Phytol. 174, 569–579. Kikuta, S.B., Richter, H., 2003. Ultrasound acoustic emissions from freezing xylem. Plant Cell Environ. 26, 383–388. Kong, W.L., Campbell, T.I., 1987. Thermal expansion due to an ice cap in an elevated water tank. Can. J. Civ. Eng. 14, 519–526. Kretschmann, D.E:, 2010. Mechanical Properties of Wood in Department of Agriculture, Forest Service Forest Products Laboratory (Eds.) Wood Handbook – Wood as an Engineering Material, General Technical Report FPL-GTR-190, Madison., pp. 100–145. Kuroda, K., Ohtani, J., Kubota, M., Fujikawa, S., 1999. Seasonal changes in the freezing behavior of xylem ray parenchyma cells in four boreal hardwood species. Cryobiology 38, 81–88. Lampriere, B.M., 2001. Ultrasound and Elastic Waves. Academic Press, New York. Lindgren, S., 1970. Thermal ice pressure. In: Proceedings of International Association for Hydraulic Research Symposium, Paper 6.7, Reykjavik. Mayr, S., Cochard, H., Ameglio, T., Kikuta, S.B., 2007. Embolism formation during freezing in the wood of Picea abies. Plant Physiol. 143, 60–67. Mayr, S., Rosner, S., 2011. Cavitation in dehydrating xylem of Picea abies: energy properties of ultrasonic emissions reflect tracheid dimensions. Tree Physiol. 31, 59–67. Mayr, S., Sperry, J., 2010. Freeze–thaw-induced embolism in Pinus contorta: centrifuge and ultrasonic experiments validate the ‘thaw-expansion hypothesis’ but conflict with ultrasonic emission data. New Phytol. 185, 1016–1024. Mayr, S., Zublasing, V., 2010. Ultrasonic emissions from conifer xylem exposed to repeated freezing. J. Plant Physiol. 167, 34–40. Muldrew, K., Acker, J.P., Elliott, J.A.W., McGann, L.E., 2004. The water to ice transition: Implications for living cells. In: Fuller, B.J., Lane, N., Benson, E.E. (Eds.), Life in the Frozen State. CRC Press, London/ Boca Raton, New York, Washington DC, pp. 67–107. Pramsohler, M., Hacker, J., Neuner, G., 2012. Freezing pattern and frost killing temperature of apple (Malus domestica) wood under controlled conditions and in nature. Tree Physiol. 32, 819–828. R Development Core Team, 2005. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, ISBN 3900051-07-0, http://www.R-project.org. Raschi, A., Scarascia Mugnozza, G., Surace, R., Valentini, R., Vazzana, C., 1989. The use of ultrasound technique to monitor freezing and thawing of water in plants. Agric. Ecosyst. Environ. 27, 411–418. Ruelland, E., Vaultier, M.N., Zachowski, A., Hurry, V., 2009. Cold signalling and cold acclimation in plants. Adv. Bot. Res. 49, 35–150. Saadat-Nia, M.A., Brancheriau, L., Gallet, P., 2011. Ultrasonic wave parameter changes during propagation through poplar and spruce reaction wood. Bioresources 6, 1172–1185. Sakai, H., Minamisawa, A., Tagaki, K., 1990. Effect of moisture content on ultrasonic velocity and attenuation in woods. Ultrasonics 28, 382–385. Salleo, S., Lo Gullo, M.A., 1986. Xylem cavitation in nodes and internodes of whole Chorisia insignis H B. Plants subjected to water stress, Relations between xylem conduit size and cavitation. Ann. Bot. 58, 431–441. Smith, A.C., Kishoni, D., 1986. Measurement of speed of sound in ice. AIAA J. 24, 1713–1715. Sparks, J.P., Campbell, G.S., Black, R.A., 2000. Liquid water content of wood tissue at temperatures below 0 ◦ C. Can. J. For. Res. 30, 624–630.

C. Guillaume et al. / Agricultural and Forest Meteorology 185 (2014) 20–25 Sparks, J.P., Campbell, G.S., Black, R.A., 2001. Water content, hydraulic conductivity, and ice formation in winter stems of Pinus contorta: a TDR case study. Oecologia 127, 468–475. Sperry, J.S., Sullivan, J.E.M., 1992. Xylem embolism in response to freeze–thaw cycles and water stress in ring-porous, diffuse-porous and conifer species. Plant Physiol. 100, 605–613. Steponkus, P.L., 1981. Responses to extreme temperatures. Cellular and subcellular bases. In: Lange, O.L., Nobel, P.S., Osmond, C.B., Ziegler, H. (Eds.), Encyclopedia of Plant Physiology. Springer Berlin, Heidelberg, New York, pp. 371–402. Teixeira, J., Luzar, A., 1999. Physics of liquid water. Structure and dynamics. In: Belissent Funel, M.C. (Ed.), Hydration Processes in Biology: Theoretical and Experimental Approaches: Nato Advanced Science Institutes Series. Series A Life Sciences 305. , pp. 35–65. Tyree, M.T., Dixon, M.A., 1983. Cavitation events in Thuja occidentalis L ultrasonic acoustic emissions from the sapwood can be measured. Plant Physiol. 72, 1094–1099.

25

Tyree, M.T., Davis, S.D., Cochard, H., 1994. Biophysical perspectives of xylem evolution – is there a tradeoff of hydraulic efficiency for vulnerability to dysfunction. IAWA J. 15, 335–360. Umbanhowar, C.E., Lambert, A.M., van Delinder, L., 2008. Effects of freezing on Young’s modulus for twigs of coniferous and deciduous trees and shrubs. Can. J. For. Res. 38, 394–399. van Dyk, H., Rice, R.W., 2005. Ultrasonic wave velocity as a moisture indicator in frozen and unfrozen lumber. For. Prod. J. 55, 68–72. Weaver, R., Lobkiss, O., 2001. Temperature dependence of ultrasonic velocity using diffuse fields: implications for measuring stress. In: Thompson, D.O., Chimenti, D.E. (Eds.), Review of Progress in Quantitative Nondestructive Evaluation, 557, pp. 1480–1486. Wisniewski, M., Lindow, S.E., Ashworth, E.N., 1997. Observations of ice nucleation and propagation in plants using infrared video thermography. Plant Physiol. 113, 327–334. Wolfe, J., Bryant, G., 2001. Cellular cryobiology: thermodynamic and mechanical effects. Int. J. Refrig. 24, 438–450.