|  |

| |

Enabling better global research outcomes in soil, plant & environmental monitoring.

Instrumentation for Total Tree Water Use and Behaviour

Only two techniques can completely account for total plant water use and behaviour: the heat ratio method (HRM); and the heat field deformation (HFD) method. This conference proceeding outlines these techniques and provides several case studies where total water use and plant behaviour have been measured in Australia, North and South America, and Europe.

There is increasing recognition in the forestry industry that water is a vital resource. Not only is water vital for tree growth but increasingly catchment management authorities are required to account for every drop of water in the landscape. Trees transpire a large volume of water however in a multi species forest certain species transpire more than others. Even in a monoculture plantation tree water use is not uniform and must be accurately measured.

Over the last 20 years there has been increasing evidence showing water movement in trees is complex. It is now recognised that water does not just move from soil to roots, stem, leaves and then atmosphere. Water can move upwards, downwards and laterally depending where there is greatest demand leaving the impression that plants exhibit “behaviour”.

This article outlines the mechanism of water movement in plants and why simple principles of physics leads to plant “behaviour”. Known as hydraulic redistribution, it is an important aspect of tree water use that must be accounted for in total tree water use. Various case studies of hydraulic redistribution are discussed that highlight the importance of this phenomenon. A number of sap flow methods are available to measure total tree water use however this article emphasises that only two of the many methods available can account for hydraulic redistribution.

Water Movement in Plants

The passage of water through soil, plants and into the atmosphere is based purely on principles of physics. Plants can have some influence over the passage of water, for example by closing stomata during periods of moisture deficit to lower transpiration rate, however they are largely passive systems influenced by environmental conditions.

A gradient in water potential is the physical mechanism that drives water movement in plants. Water potential is the amount of work, or energy, available when compared with a reference condition. The simplest analogy is a ball that is sitting at the crest of a hill. At this point the ball has a large potential to do work – i.e. roll down the hill; if the ball was placed at the bottom of the hill it has low potential to do work – i.e. there is little energy available for the ball to move up the hill. Similarly, water moves from areas of high potential to low potential; water always moves downhill and never uphill.

Pure water has the largest potential to do work and is equivalent to a ball at the crest of a hill. At this point a value of zero is given. The SI unit for water potential is pascal (Pa) therefore water at its reference point has a value of 0Pa (many other terms are also used, interchangeably, such as bar, joules, relative humidity, and also kilopascal and megapascal – here we will use megapascal, MPa). The gradient of water potential becomes more negative from the reference point of 0MPa . At a value of -0.001MPa there is a small gradient in water potential, the equivalent of rolling the ball down a slight decline. At a value of -100MPa there is a large gradient in water potential, the equivalent of rolling the ball off a sharp cliff.

Values found in soil that are important to plants range from -0.033MPa (field capacity) to -1.5MPa (wilting point). Leaf water potential can be as low as -3.0MPa for a drought stressed plant. Air dry, or atmospheric, values are in the order of -100MPa (or 47.76% relative humidity). There is a strong water potential gradient through the soil-plant-atmosphere continuum from -1.5MPa to -3.0MPa to -100MPa (for very dry conditions). This is the physical principle that allows water movement through plants and ultimately for transpiration to take place.

Over the last 20 years it has been increasingly recognised that not only does the water potential gradient allow moisture to move from soil to the atmosphere, but it allows water to move from any part of the plant to any other part of the plant. There is now a large amount of evidence documenting water movement down stems, across stems, from stems to roots, and across various parts of the root profile. The term is hydraulic redistribution and it is an extremely important mechanism that allows plants to cope with moisture deficits.

Case Studies of Hydraulic Redistribution in Trees

Stem-mediated Hydraulic Redistribution

Douglas-fir (Pseudotsuga menziesii) is a coniferous tree in the family Pinaceae that can tolerate low rainfall environments. Nadezhdina et al. (2009) were interested in how a tree growing in a low rainfall zone (560mm rainfall per annum) coped with moisture deficit. A representative tree from a mixed forest experimental stand (0.5ha) was selected for experimentation. The tree was 53 years old and 35.6cm in diameter at breast height.

Sap flow instrumentation (heat field deformation, HFD, method) consisted of four sensors installed in the tree. Two sensors were installed in the stem (north face and south face) and two sensors installed in large roots at least 16cm from the trunk (northern root and southern root). Monitoring of sap flow began during dry conditions and then the researchers irrigated the southern side of the tree only. In this exploratory experiment, the researchers closely monitored sap flow at various locations around the tree as well as at various depths of sapwood within each location.

Nadezhdina et al. found a strong pattern of hydraulic redistribution in this Douglas-fir. Water was transported from irrigated soil, through the root system on the southern side of the tree, towards the south facing stem. Water was then transported across the trunk to the north facing stem, then in reverse flow down to the root system on the northern side of the tree. The researchers did not have any instrumentation to account for water potential gradients in this process and it would have been interesting to see results from a stem psychrometer or soil water potential sensor. Nevertheless, they concluded that due to water potential gradient water moved from an irrigated portion of the tree to the non-irrigated portion.

Foliage-mediated Hydraulic Redistribution

The most common condition for plants is leaf water potential to be lower (more negative) than stem water potential – a condition necessary for the passage of moisture from roots to the atmosphere. However there are certain environmental conditions that allow for leaf water potential to be higher than stem water potential allowing for the movement of moisture from leaves to the stem. Burgess and Dawson (2004) hypothesised that moisture may even travel from the atmosphere to the leaves, stem and roots; a remarkable condition of reverse sap flow where the soil-plant-atmosphere continuum challenges conventional theory.

Sap flow instrumentation in this study was based on the heat ratio method (HRM) as it has the ability to detect low and reverse flow rates (Burgess et al . 2000). The researchers studied a number of trees however the installation for a single tree consisted of three sets of probes at breast height, two at approximately 50m (trees were 60 to 70m in height), and six sets of probes placed in three separate branches high in the tree (a probe on the lower side of the branch and another on the upper side).

Burgess and Dawson found a number of interesting sap flow patterns. Nocturnal sap flow was observed during a night of low relative humidity (between 20 and 40%) supporting the notion that redwood has porous stomata. On Day 1 a typical diurnal course of sap flow was observed with maximum transpiration rates around midday. On Day 2, however, there was heavy fog and transpiration ceased altogether. Instead, reverse sap flow was observed at branch, 50m and breast height sensors. The rate of reverse flow was as high as 7% of the previous day’s transpiration. The mechanism by which moisture enters the leaf is via the hyphae or hairs that extend from stomata that act as wicks to draw the water back in.

The authors concluded that they observed a clear pattern of reverse sap flow and water uptake by the leaves. Water moved from the leaves, into branches, stem and, possibly (it was not measured), roots and soil. This observation is considered to be an adaptation to moisture deficits experienced by the coast redwood. Additionally, it is an important component of the water balance of this forest that needs to be taken into account in hydrological models.

Root-mediated Hydraulic Redistribution

The redistribution of moisture by roots is the most widely documented pattern of hydraulic redistribution. Initially, the pattern was discovered by using soil moisture sensors (Richards and Caldwell 1987; Caldwell and Richards 1989). Termed hydraulic lift, roots deep in the soil profile with access to moisture transport water to roots in the shallow soil profile where the soil is drier. The physics behind the process is a water potential gradient from moist to dry soil. Not only does hydraulic lift contribute to the water balance of the entire root system, but roots in shallow soil are able to maintain access to higher nutrient levels. Root hydraulic lift is only observed at night as the daytime water potential gradient is far stronger than any gradient that can be produced in the soil.

As hydraulic lift became more widely documented, more interesting research questions were posed. For example, in certain parts of the Amazon Basin there is a strong wet-dry season. However trees of the luxuriant rainforest rarely show signs of drought stress towards the end of the dry season. Oliveira et al. (2005) suspected hydraulic lift enabled trees to maintain a favourable water status during the dry season. Further, Oliveira et al. hypothesised that at the onset of the wet season soil moisture would be redistributed from wetter shallow soil to drier deep soil via the root system. This pattern of hydraulic redistribution would be particularly apparent as this study was conducted on heavy clay soil where the infiltration of rainfall would be slow.

Sap flow instrumentation was again based on the heat ratio method (HRM) for its ability to measure reverse flow (Burgess et al . 2000). The experiment was carried out at the Floresta Nacional do Tapajo ́s, Brazil, where control and treatment trees were established. The treatment plot consisted of a “rainout” where a shelter was erected above the soil surface to exclude rainfall. Soil was excavated down to 1m and sap flow sensors installed on all tap roots and up to four lateral roots.

During the dry season, there was reverse nocturnal sap flow in the roots indicating moisture was moving from roots into the soil – a pattern consistent with hydraulic lift. With the onset of wet season rainfall there was positive nocturnal sap flow in the lateral roots yet there was still reverse flow in the tap root indicating moisture movement from wet top soil, into lateral roots, towards the tap root and down into deeper soil profile. This pattern continued for approximately seven days into t he wet season. In the treatment plot hydraulic lift was observed at all times.

With changes in climate and rainfall patterns it is important for forestry to understand how plants respond to moisture gradients. Results from studies on hydraulic redistribution clearly demonstrate that any water balance model, particularly catchment models, must account for reverse sap flow. Knowledge of the total amount of water use by trees is still critically important.

Case Studies of Total Tree Water Use

Phytoremediation and Groundwater Monitoring at Kamarooka

Salinity is a major economic and environmental problem, particularly in Australia. In north central Victoria, north of Bendigo, salinity has had a large impact on the landscape. In 2003 the Northern United Forestry Group (NUFG) decided to undertake a project to reclaim a barren land and in the process restoring functional ecosystem processes and increasing agricultural production (Figure 1). There was considerable interest in establishing salt tolerant trees in order to provide income from forestry projects and provide habitat for fauna. Sugar gum (Eucalyptus cladocalyx), flat-top yate (E. occidentalis), willow wattle (Acacia salicina) and Eumong (A. stenophylla) were planted at the saline site, Kamarooka. Fourteen monitoring bore holes were established under the trees and in non-forested areas to monitor groundwater.

Throughout 2007 a 3m groundwater depression was formed under the forested area. To ascertain whether the trees were the causal effect behind the lowering of the groundwater it was essential to established total tree water use. Sugar gum and flat-top yate were instrumented with sap flow sensors using the heat ratio method (HRM). Sensors were installed in mid-January 2008 and monitoring commenced.

Between April and September 2008 average tree water use was 5 litres per day. On a plantation scale tree water use was approximately 25,000 litres per day per hectare. At the start of this period the groundwater was at a depth of approximately 4m and by March 2010 it had dropped to approximately 6m – in spite of a few heavy rainfall events.

Monitoring sap flow and total tree water use at Kamarooka has clearly demonstrated that trees can be used as a phytoremediation tool on saline land. Forestry can be established to provide income and environmental benefits. For more information please visit the NUFG website: http://nufg.org.au/index.html


Figure 1. Changes in the landscape at Kamarooka, north of Bendigo, Victoria, Australia. Flooded saline land, the establishment of a plantation, and the growth of trees a few years after planting.

Establishing a Forest in a Desert:

Antamina Mine, Peru One of the world’s largest copper and zinc mines is Antamina located in the Andes of Peru. It is a multi-billion dollar investment located 4,300m above sea-level. Copper and zinc is gravitationally transported in slurry via a 300km long pipe to a seaport. At the seaport copper and zinc is separated from the slurry. The mine operators then had a major problem to overcome. The leftover water could be dumped into the ocean at considerable environmental cost, a filtration plant could make the water fit for agriculture or human consumption at considerable economic cost, or a plantation could be established and irrigated with the leftover water.

The trees would act as “biopumps”, transpiring the leftover water into the atmosphere and preventing through drainage of contaminated water into the underlying aquifers that provide much of the potable water for the surrounding population. This last option was additionally attractive as the area around the seaport is desert and it would be an afforestation project (rather than displacing existing forest or agricultural land).

A 174 acre forest was established with 190,000 individual trees consisting of eight species (Figure 2) and mine production continued. It quickly became obvious that tree transpiration was not uniform throughout the year and changed with season. In order to avoid flooding the forest during periods of low transpiration (and thereby risking through drainage and pollution of aquifers) it was necessary to know how much water the trees transpired. Moreover, the mine operators wanted to increase production which entailed increasing irrigation and forest area to maintain equilibrium with effluent input to transpired output. The most efficient method of removing the extra leftover water was to plant trees that transpired the most amount of water.

Sap flow sensors employing the heat ratio method (HRM) were installed on nine trees of three species: Acacia spp, Tamarix spp, and Algorrobo spp. The results over an 18 month period showed Acacia, Tamarix and Algorrobo transpired approximately 35,000, 16,500 and 6,000 litres of water respectively.

Quantifying total tree water use allowed more precise management decisions to be made in terms of adherence to environmental regulations plus, the data to accurately expand the mines capacity whilst continuing to meet current and potentially future modifications to environmental regulations.


Figure 2.The use of leftover water from a mining operation allowed the establishment of a forest in the desert.

Sap Flow Instrumentation

Heat Ratio Method (HRM)
HRM was initially developed by Burgess et al. (2001) and is now available on a commercial basis through ICT International. HRM sensors consist of three needles (Figure 3A): a central heater needle and two sampling needles inserted upstream and downstream of the heater. The needles are 3.5cm in length with a thermocouple located at 1.25cm and 2.75cm (called outer and inner measurement points, respectively). Research has demonstrated sensors located at these two depths account for the vast proportion of sap flow in trees (Ford et al. 2004; Fiora and Cescatti 2006). Figure 3B outlines sensor design and the location of thermocouples in sapwood.
The HRM measures the ratio of the increase in temperature, following the release of a pulse of heat, at points equidistant downstream and upstream from a line heater. In Figure 3B, temperature sensors “a” and “b” are linked as are sensors “c” and “d”. The ratio of increase in temperature between measurement points “a” and “b” indicate sap velocity as well as sap direction. A low ratio indicates little to no movement in sap (i.e. the tree is not transpiring) whereas a large ratio indicates rapid sap flow (i.e. it is a hot, sunny day). Because the measurement points are located equidistant from the central heater it is then possible to determine direction of flow due to the polarisation of the heat flux: if the value is positive then sap is travelling up the stem; if the value is negative then there is reverse sap flow.
Figure 3. The heat ratio method (HRM) consists of three needles connected to a stand-alone datalogger (A) with a central heater and two temperature sensing needles, each with two measurement depths spaced radially across the sapwood (B).

Heat Field Deformation (HFD)

The HFD technique is a radically new method for measuring sap flow. It is ideally suited to sap flow research projects that require the measurement of extended radial sap flow profiles to accurately map hydraulic architecture of trees. Similar to the HRM sap flow sensor the HFD meter can measure high sap flow rates as well as low to zero and reverse sap flow. Hence as both sensors can measure in the same range the HFD technique provides an extension of the HRM method making both sensors highly complementary to each other in most sap flow measurement applications. Developed by Professor Nadezda Nadezhdina, Mendel University, Czech Republic the HFD technique has been used in published sap flow research since 1998 to study many previously unanswered plant physiological questions (Nadezhdina et al. 1998). HFD instruments are commercially available through ICT International.
The HFD technique is a thermodynamic method based on measuring the difference in temperature (dT) of the sapwood both symmetrically (in the axial direction, above and below) and asymmetrically (in the tangential direction or to the side) around a line heater (Figure 4A).
The heater is continuously heated and generates an elliptical heat field under zero flow conditions. Sap flow significantly deforms the heat field by elongating the ellipse (Figure 4B). The symmetrical temperature difference (dT sym) allows bi-directional (acropetal and basipetal) and very low flow measurements, whereas asymmetrical temperature difference (dT as) is primarily responsible for the magnitude of medium and high sap flow rates.
By using the ratio of measured temperature differences and applying correction for each measurement points local conditions, using the adjustable K-values, the common features of the medium (such as variable water content, natural temperature gradients and, wound effects) have negligible impact on sap flow calculations. The value for parameter K is equal to the absolute value of dT sym or dT as for a zero flow condition. Under flow conditions the parameter K can be extrapolated with accuracy using linear regression.
Figure 4. The heat field deformation (HFD) method consists of four three needles (A). The heater needle has a constant source of heat which the movement of sap within the stem deforms (B) and is able to be detected by temperature sensors.

Sap Flow Tool Software

Accurate calculation of corrected sap flow based on raw sap velocities, thermal properties of wood, wound effects of inserting needles, as well as differing values of sapwood density amongst species, is laborious and time consuming. A software package, Sap Flow Tool, has been developed that calculates corrected sap flow based on input parameters. The software package is designed to support raw sap flow data derived from HRM and HFD instruments. More information can be found at: http://www.sapflowtool.com
Figure 5 demonstrates the valuable information of hydraulic architecture derived from the HFD technique. Figure 5A shows sap flux densities for a single day for five measurement points located at depths 0.3cm, 0.9cm, 1.5cm, 2.1cm and 2.7cm. The vertical line indicates 9 AM or pre-dawn on this particular day. The lower half of the graph is the radial profile and shows not sap movement. Figure 5B is sap flux densities for the same five measurement points on the same day. The vertical line is now at 3 PM – the approximate time of maximum sap flow on this particular day. The lower portion of the graph now shows sap movement across the radial profile. There is a slight increase around the second measurement point, located at a radial depth of 0.9cm across the sapwood. Throughout the day the highest rates of sap flow are seen at the second and third measurement point with some increase at the fourth. This indicates that for this particular tree sap flow is centred on a radial depth between 0.75cm and 2.25cm.
Figure 6 is a 3D visualisation on sap flow across the radial profile over a four day period. It is clear from this graph that although sapwood depth is 8cm for this tree sap flow is centred on a radial depth of 0.75cm to 2.25cm depth.


Accurate measurement of sap flow in trees is extremely important for forestry. Not only is it important to account for total tree water use but to elucidate pattern s of hydraulic activity (“behaviour”) within trees. Over the last 20 years such research has uncovered unexpected results and led to the term hydraulic redistribution. Knowing the hydraulic behaviour of plant s is important in a future world of climate change, variable rainfall, and water use accountability. Having a suite of quality tools to enable accurate measurement of parameters such as total daily water use rates provide managers with the data to make informed management decisions both operationally as well as at planning and policy levels. These decisions will impact the use and management of our forest resource both now and in the future.
Figure 5. Sap velocity on a representative day graphed in Sap Flow Tool software and data collected from an HFD instrument. (A) The vertical line indicates sap velocity at 9 AM which is zero flow across the entire profile. Zero flow can be seen in the lower half of the graph. (B) The vertical line indicates sap velocity at 3 PM which is at its maximum rate for this particular day. The different lines on the graph are derived from 5 different measurement points located along the HFD needle. The lower half indicates that sap flux density in this radial profile is highest at the second (0.9cm sapwood depth) and third measurement point (1.5cm depth) and minimal at the first and fifth measurement points. Sap flow in this profile is mostly undertaken between 0.75cm and 2.25cm depths radially across the sapwood.
Figure 6. Sap flow over a representative four days graphed in Sap Flow Tool software and data collected from an HFD instrument. 3D representation of data easily depicts the location in sapwood where water movement is taking place.


Burgess SSO, Adams MA, Bleby TM. 2000. Measurement of sap flow in roots of woody plants: a commentary. Tree Physiology 20: 909-913.
Burgess SSO, Adams MA, Turner NC, Beverly CR, Ong CK, Khan AAH, Bleby TM. 2001. An improved heat pulse method to measure low and reverse rates of sap flow in woody plants. Tree Physiology 21: 589-598.
Burgess SSO, Dawson TE. 2004. The contribution of fog to the water relations of Sequoia sempervirens (D. Don): foliar uptake and prevention of dehydration. Plant, Cell and Environment 27: 1023-1034.
Caldwell MM, Richards JH. 1989. Hydraulic lift: water efflux from upper roots improves effectiveness of water uptake by deep roots. Oecologia 79: 1-5.
Fiora A, Cescatti A. 2006. Diurnal and seasonal variability in radial distribution of sap flux density: implications for estimating stand transpiration. Tree Physiology 26: 1217-1225.
Ford CR, McGuire MA, Mitchell RJ, Teskey RO. 2004. Assessing variation in the radial profile of sap flux density in Pinus species and its effect on daily water use. Tree Physiology 24: 241-249.
Nadezhdina N, Cermak J, Nadezhdin V. 1998. Heat field deformation method for sap flow measurements. In Measuring Sap Flow in Intact Plants , Cermak J, Nadezhdina N (eds). Proceedings of the 4th International Workshop , Zidlochovice, Czech Republic, IUFRO Publications: Brno; 72-92.
Nadezhdina N, David TS, David JS, Ferreira MI, Dohnal M, Tesar M, Gartner K, Leitgeb E, Nadezhdin V, Cermak J, Jimenez MS, Morales D. 2010. Trees never rest: the multiple facets of hydraulic redistribution. Ecohydrology DOI: 10.1002/eco.148.
Nadezhdina N, Steppe K, De Pauw DJW, Bequet R, Cermak J, Ceulemans R. 2009. Stem- mediated hydraulic redistribution in large roots on opposing sides of a Douglas-fir tree following localized irrigation. New Phytologist 184: 932-943.
Oliveira RS, Dawson TE, Burgess SSO, Nepstad DC. 2005. Hydraulic redistribution in three Amazonian trees. Oecologia 145: 354-363.
Richards JH, Caldwell MM. 1987. Hydraulic lift: substantial nocturnal water transport between soil layers by Artemisia tridentata roots. Oecologia 73: 486-489.