By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Due to essential maintenance the subscribe/renew pages will be unavailable on Wednesday 26 October between 02:00- 08:00 BST/ 09:00 – 15:00 SGT/ 21:00- 03:00 EDT. Apologies for the inconvenience.
Diameter variations in the xylem and whole stem (i.e. over bark) stem of a Scots pine (Pinus sylvestris L.) tree were measured at four heights over a 23 d period at 5 min intervals. Cross-correlation analysis was used to calculate time lags between the measurements. Xylem diameter measurements at the different heights had time lags varying from 10 to 50 min, measurements at the lower heights lagging behind the most. This result was in good agreement with the cohesion theory of transpiration. For the whole stem diameter measurements, the treetop lagged behind all other heights and the shortest lags were midway along the stem. Changes in whole stem diameter always lagged behind those of xylem stem diameter (30–110 min), and at all heights. The considerable differences in the behaviour of xylem and whole stem diameter support the Münch hypothesis of phloem flow. Time lags calculated separately for the shrinkage (morning) and swelling (afternoon) periods indicated shorter time lags during the swelling periods. The non-destructive methods used show promise in the simultaneous study of flow dynamics of xylem and phloem in trees.
Tree stem diameter changes are known to consist of two components: an irreversible component due to growth and a variable component due to moisture content that shows a diurnal cycle. Measurement of these diurnal diameter changes provides information about sap flow and the water tension inside the stem (Zimmermann & Milburn 1982; Zimmerman 1983). However, measuring over-bark diameters does not provide information directly about the water flow in the xylem because some of the change in diameter is due to the shrinkage and swelling in phloem and bark cells. Some studies even suggest that all or almost all of the diurnal changes in diameter are due to changes in the outer tissue (Dobbs & Scott 1971; Hinckley & Bruckerhoff 1975, Wronski, Holmes & Turner 1985). In contrast, Hellkvist, Hillerdal-Hagströmer & Mattson-Djos (1980) showed that the amplitude of xylem diameter changes to be about half that of the whole stem diameter changes for Scots pine (Pinus sylvestris L.) and Neher (1993) that the xylem contributes about 30% to the diameter variation of Monterey pine (Pinus radiata).
Sap flows from the roots to the leaves in the xylem and sugars are transported from the leaves to the growing parts or reservoirs in the phloem. The sap flow direction in the xylem is upwards and the flow in the phloem is mostly downwards. However, carbohydrates are transported to the regions of greatest demand and, depending on the stage of growth, there may be an upward flow in the phloem (Kramer & Kozlowski 1979). Besides the vertical movement, water (and solutes) also moves radially between phloem and xylem along a decreasing water potential gradient (Nobel 1991).
Measuring sap flow in field conditions is not easy because, for example, the equipment tends to break the cell structure, which may lead to embolization and reduction of conductance near the measuring system. However, there are some reliable, indirect and non-destructive methods to measure sap flow in the xylem (Smith & Allen 1996), but measuring the flow in the phloem is even more problematic. Phloem is fairly thin and there is considerable interaction with the surrounding cells (Nobel 1991; Peuke et al. 2001).
All the measurements were made at the Helsinki University SMEAR II field measurement station (Station for Measuring Forest Ecosystem–Atmosphere Relations) in Hyytiälä, southern Finland (61°51′ N, 24°17′ E, 181 m a.s.l.) (Vesala et al. 1998). The station is located in a homogenous 38-year-old (in 2000) Scots pine (Pinus sylvestris L.)-dominated stand. The height of the dominant trees was 13 m and the mean diameter at breast height (1·3 m) was 13 cm. The total (all-sided) leaf area index varied from 4 to 6, the maximum occurring in late August. The stem biomass is 41 t ha−1 and the needle biomass 5·1 t ha−1 (Ilvesniemi & Liu 2001).
The stem diameter variation of a 13·3-m-high pine tree was measured every 5 min using eight linear variable displacement transducers (LVDT; model AX/5·0/5; Solartron Inc., Bognor Regis, West Sussex, UK) at several heights for a 23 d period, 20 June to 12 July 2000. Each transducer was attached to a rigid steel frame mounted on the tree with attachment plates 15 cm above the measurement point. The transducers were arranged in four pairs at four different heights (2·1, 7·4, 10·4 and 12·0 m) (Fig. 1). The vertical distance between sensors in a pair was about 20 cm; the lower transducer measured the xylem diameter and the upper sensor measured the whole stem diameter. For the xylem diameter measurements an area of about 1 cm2 of phloem, bark and cambium was removed on opposite sides of the tree and covered with aluminium plates. One end of the frame was set to rest on one plate and the head of the sensor positioned on the other. For the whole stem diameter measurements, the bark was smoothed before attaching the aluminium plates. Each frame and transducer was covered with a conical polyethylene shade to avoid heating by direct radiation. The frame temperatures were measured using copper–constantan thermo-couples and used to correct for the effect of temperature on the expansion of the frame (Sandvik 1802 Steel, Sandvik, Sandvixen, Sweden). Irvine & Grace (1997), Vesala et al. (2000) and Perämäki et al. (2001) have used similar installations in their studies. The photosynthetic photon flux density (PPFD) above the canopy (15 m) was measured with a quantum sensor (LI-190SA; Li-Cor Inc., Lincoln, NB, USA) attached to a separate tower. Dew point and air temperatures, used to calculate relative humidity, were measured at a height of 23 m using a chilled mirror sensor (M4 Dew point monitor; General Eastern, Woburn, MA, USA) and a resistance thermometer (Type Pt-100), respectively.
The sap flow in two branches was measured using the heat balance method (Dynamax Flow32®; Dynamax Inc., Houston, TX, USA). The stem surface was smoothed and a thin layer of silicone grease applied to the surface before fitting the gauges. The installation and selection of measurement points was carried out as recommended by the manufacturer (Dynamax 1990). The branches forked from the main stem at heights of 8 and 9 m. The diameters were 2·1 and 1·0 cm, leaf areas 1·60 and 0·97 m2 and lengths 1·6 and 1·4 m, respectively. The branch at 8 m was oriented south-west and the branch at 9 m south-east. The leaf areas of the branches were estimated from the regression between the branch basal sapwood area and the needle mass as determined for the same stand (Palmroth, Berninger & Nikinmaa 1997).
The changes in diameter followed a similar pattern to that reported earlier (e.g. Hellkvist et al. 1980; Sevanto et al. 2001), being the largest in the daytime, particularly on clear warm days, when sap flow was also the highest (Fig. 2). The amplitude of the diurnal cycle was about a tenth of a millimetre at its maximum. Growth was seen only in the whole stem measurements. During the measurement period the growth was linear, being the largest at the treetop (1·6 mm in 23 d) and smallest at the base (0·5 mm in 23 d). The linear trend was subtracted from the whole stem diameter change measurements to reveal the diurnal change in diameter. The mean diurnal amplitude in whole stem diameter was 2·6 ± 0·4 times that of the xylem diameter, and indicated that almost one-third of the whole stem diameter change was due to changes in xylem and two-thirds to the living tissue outside the xylem.
In absolute values, the xylem diameter variation was largest at a height of 2 m and decreased with increasing height (Fig. 2c). The amplitude of the whole stem measurements did not vary significantly with respect to height (Fig. 2d). In order to compare the changes in diameter at different heights the changes in diameter were normalized by dividing the values by the equivalent stem diameter at the corresponding height, the sapwood diameter in the case of the change in xylem diameter measurements. The thickness of the tissue outside the xylem (cambium, phloem and bark) varied from 2 mm at the top of the tree to 3 mm at the base. In contrast to the absolute values, the amplitude of the normalized xylem diameter values did not vary much with height but that of the whole stem measurements increased slightly towards the treetop (Fig. 3).
Time lags were calculated by finding the highest cross-correlation between diameter measurements at any two heights when one was moved in time. Cross-correlation values were calculated for each day separately and averaged over the measurement period for each sensor. The statistical significance of the differences between the values was tested using a t-test. This assumes that the data for separate days are independent, which may not be the case. Over the 23 d measurement period the average time lags for the xylem diameter measurements varied from 10 to 50 min, the lag increasing with increasing distance from the treetop (Fig. 4a). The apparent inconsistency of the xylem time lags (the sum of the lags of adjacent measurement heights does not equal the largest time lag) results from the averaging and rounding of the values.
The time lag for the whole stem diameter measurements were slightly shorter, varying from 5 to 40 min, but the order of the increasing time difference did not vary regularly with height. The uppermost measurement was behind all the others and the lowest was behind the two in the middle , which were almost in phase (the second highest was 5 min behind). The time response of the second lowest sensor (height 7·40 m, 40 cm above the first living branch) was the most rapid of all. A comparison of xylem and stem measurements in each sensor pair showed that at every level the stem measurement was behind the xylem one. The shortest lags (30 min) were at the two lowest levels and the largest (110 min) at the top of the tree.
Time lags during the shrinkage period (from 0400 to 1200 h) (Fig. 4b) were generally somewhat shorter than those calculated over the whole day, but the differences were not statistically significant (P > 0·1). Swelling periods (from 1300 to 2400 h) showed even shorter time lags (Fig. 4c) and they statistically differed from both the shrinkage period and whole day data (P < 0·01) in almost all the cases. Only the differences between the time lags of the three highest xylem measurements were not significant.
We also compared time lags on days of high PPFD (maximum PPFD > 1200 µmol m−2 s−1) to those with moderate (PPFD < 1200 µmol m−2 s−1) and found no statistically significant difference using whole day, shrinkage or swelling period values. Neither were there any significant differences in the time lags between days with moderate and low relative humidity (moderate: relative humidity > 50% for whole day, low: minimum relative humidity < 50%).
According to the cohesion theory of water movement in plants, transpiration should create a tension gradient that pulls water up inside the xylem. Since wood material is elastic, varying tension is reflected in the stem diameter (Irvine & Grace 1997; Perämäki et al. 2001). Xylem is more rigid material than sapwood and contracts less than elastic living tissue, as our results indicate. However, the large amplitude of the whole stem diameter variations may partly result from water exchange between the xylem and the outer tissue. The cells of the phloem, cambium and bark may serve as a store of water. When transpiration is highest, water is taken from the store most actively and during the swelling period the store is refilled (Herzog et al. 1995). The similarity of the amplitudes of the whole stem shrinkage at all heights (Fig. 2d) results from comparable bark and phloem thickness along the stem. The normalized amplitudes (Fig. 3) reflect the decrease in sapwood area and invariability of phloem and bark thickness with height. On some occasions, the diameter changes at height S3 (Fig. 2d) exhibited smaller amplitudes than at other heights, but the difference was small and it may have only been due to local inconsistencies in stem structure.
The focus of this study was on the timing of the diurnal diameter change along a tree stem and between the xylem and the outer tissue, and its relation to sap flow dynamics. Theoretically, the maximum speed of a pressure signal inside some material corresponds to the sound velocity, which, in this case, would give an instantaneous response. The lower limit to the speed of signal propagation is the actual stream velocity, which in this case was of an order of centimetres per hour. The time lags of xylem diameter change were found to increase with increasing distance from the treetop, the lower parts of the stem being behind in most cases. The magnitude of the lags was of the order of 10 min, which indicates that there are some mechanisms slowing down the theoretically instantaneous signal propagation. Traditionally, time lags of stem diameter changes have been related to the storage capacity of the tissue along the pathway (Whitehead & Jarvis 1981; Milne et al. 1983; Wronski et al. 1985; Herzog et al. 1995 and Zweifel, Item & Häsler 2000). However, the elastic properties of wood material would also contribute to signal propagation and affect the timing of diameter changes. Taking this into account, the values presented support the cohesion theory, i.e. similar dynamics of diurnal diameter changes are obtained when the tension, created by transpiration, are propagated through the entire stem (Perämäki et al. 2001). The magnitude of the lags on one hand and the similarity between the lags of the shrinkage and swelling periods and those of the whole day on the other hand suggest that storage in the xylem cannot be large.
Comparing the diameter variation of the xylem and the whole stem at the same height revealed that the whole stem diameter change always lagged behind. The lags were fairly large (from 30 to 110 min) and the largest lag was observed at the treetop. Whitehead & Jarvis (1981) consider that water is first taken from stores closest to the origin of water loss. In order for the water stored in the phloem and bark tissue to move easily to the transpiring surface would require that the water is hydraulically connected to the water in the xylem (Zweifel & Häsler 2001). However, the longer lag we observed at the treetop would suggest that the hydraulic conductivity between the store and the transpiration stream close to the transpiring surfaces was low. Furthermore, considering that wood permeability decreases towards the distal parts of the tree in Scots pine (Zimmermann 1983), the hydraulic connection between the store and the transpiration stream would be expected to have a high resistance. Our results therefore do not support the importance of phloem and bark as storage (see also Whitehead & Jarvis 1981).
The observed time lags and diameter variation may also be related to the transport of carbohydrates in the phloem. Since the change in xylem diameter is a minor part of the total stem diameter change and only a thin layer of bark was left to cover the phloem, the stem diameter changes are considered to mainly represent changes in the phloem. According to the Münch hypothesis, phloem flow is driven by a turgor pressure gradient caused by the loading and unloading of sugars into the phloem sap (Kramer et al. 1979). The loading of sugars at the treetop lowers the osmotic potential in the phloem and results in a water flow from the xylem to the phloem. This increases the phloem volume but the net effect on the stem diameter remains small. The result would be a large time lag between the xylem and the stem measurements at the top of the tree. Lower down in the stem, the growing meristem uses sugars and the osmotic potential increases. Water would readily flow into the xylem as the tension gradient builds up due to transpiration. This in turn would lower the turgor pressure in the phloem and decrease the time lag between the xylem and the whole stem shrinkage lower down the stem (note that the difference between the time lags of the two lowest measurement points is only 5 min and it is not statistically significant). Thus our measurements suggest that the time lags between the xylem and the whole stem shrinkage reflect changes in turgor pressure in the phloem, as implied by the Münch hypothesis of phloem flow.
This study also revealed some problems in the use of cross-correlation in time lag analysis. Data with periods when the diameter decreased or increased quite linearly, the noise in the measurements affected the correlation coefficient relatively more than the overall trend and the determination of the actual time lags was difficult. This problem was most evident with afternoon and cloudy-day values, although the periods were selected to include at least one of the daily maximum or minimum points of the diameter change curve. For this reason, the two negative xylem time lags for swelling periods seen in Fig. 4c should be viewed with caution. Nevertheless, this study showed that time lag studies is a useful tool in analysing the flow dynamics within a tree. More detailed and longer-lasting experiments are required.