Protocol

Authors

Cleiton B. Eller, Stephen S.O. Burgess, Rafael S. Oliveira

Overview

This protocol explains how to set up a sensor to measure sap flow using the heat-ratio method. We also explain how to convert the raw heat-pulse velocity in sap flow data.

Background

The Heat-Ratio Method (HRM; Burgess et al. 2001) is a modification of another popular Heat-Pulse Velocity based method to measure sap flow, the Compensation Heat-Pulse Method (CHPM; Swanson and Whitfield 1981, Green and Clothier 1988). The CHPM has been successfully used to measure sap flow in roots and stems in several studies (Becker 1996; Lott et al. 1996; Zotz et al 1997; Burgess et al. 1998). However, CHPM is inaccurate at low sap flow rates (e.g.  < 3.6 ± 7.2 cm h^-1; Becker 1998) and incapable of measuring reverse sap flow.

The HRM is based on the temperature ratio created between probes positioned symmetrically above and below a line heater. The HRM successfully measures low sap flow rates, being widely used to study sap flow in roots (Burgess et al 1998), nighttime sap flow (Dawson et al 2007; Rosado et al 2012) and during any conditions of low atmospheric demand (Burgess & Dawson 2004). Besides that, the method is capable of accurately measuring reverse-flow, which makes it suitable to study water fluxes resulting from hydraulic redistribution (Burgess et al 1998, Oliveira et al 2005, Burgess and Bleby 2006) and foliar water uptake (Burgess & Dawson 2004, Goldsmith et al 2013; Eller et al 2013). Generally, it is assumed that HRM have a theoretical upper limit to measure sap velocities of 54 cm hr^-1 (Swanson 1983; Burgess et al 2001). However, laboratory tests using a high-pressure pump and cut stem segments suggests that HRM could measure accurately sap flow ranging from 0-100 cm hr^-1 (Burgess and Downey 2014). Model simulations from Vandegehuchte et al (2014) also show that HRM provides accurate results even at high sap velocities (up to 100 cm hr^-1) when stem temperature gradients are not present.

Materials/Equipment

• Thermometer probes, heater probe and data logger (or a commercially available HRM device)
• Drill and drill bits
• Drill guide
• Increment borer
• Bark depth gauge
• Methyl Orange solution (0.1%)
• Measurement tape
• Chisel and mallet

Units, terms, definitions

We follow the unifying nomenclature for sap flow measurements given in Edwards et al (1996). The other units and nomenclature follow Burgess et al (2001).

Heat pulse velocity (v_h, cm hr^-1)

Heat pulse velocity corrected for wound effects (v_c, cm hr^-1)

Sap velocity (v_s, cm hr^-1)

Sap flow (Q, cm³ hr^-1)

Temperature increases in downstream and upstream probes after the heat-pulse (v₁ and v₂, respectively, °C)

Temperature before and after the heat pulse (T₀ and T_h, respectively, °C)

Sapwood thermal diffusivity (k, cm² s^-1)

Fresh sapwood thermal conductivity (K_gw. J m⁻¹ s^-1 °C ⁻¹)

Dry sapwood matrix thermal conductivity (K_w, J m^–1 s ^–1 °C^–1)

Fresh sapwood specific heat capacity (c, J kg^–1 °C^–1)

Dry sapwood specific heat capacity (c_w, J kg^–1 °C^–1)

Sapwood density (ρ_b, kg m^-3)

Sapwood specific gravity (G, kg_{dry wood} (m³_wood)⁻¹ (kg_water (m³_water)⁻¹)⁻¹

Sapwood water content (m_c, kg)

Sapwood water content at fiber saturation point (m_{c_FSP}, kg)

Sapwood void fraction at the fiber saturation point (F_{v_FSP}, m³_void m^-3_wood)

Sapwood dry weight (w_d, kg)

Sapwood fresh weight (w_f, kg)

Cell wall density (ρ_cw, kg m^-3)

Water thermal conductivity (K_s, J m^–1 s ^–1 °C^–1)

Water specific heat capacity (c_s, J kg^–1 °C^–1)

Water density (ρ_s, kg m^-3)

Wound correction coefficient (B)

Total xylem cross-sectional area (CSA, cm²)

Procedure

The protocol for tree selection and other general sap flow installation recommendations are described here. We will discuss only the procedures that are particular to HRM in the sections below.

 

Probes positioning

• In HRM, two thermometer probes are positioned equidistantly from a probe equipped with a heater. In the original HRM principle, the distance between the probes and the heater was 6 mm. Nowadays, the most used distance is 5 mm, as this increases the maximum (theoretical) measurable sap velocity rates from 45 cm hr^-1 to 54 cm hr^-1 (Swanson 1983; Burgess et al 2001).
• The thermometers should be positioned on the most active xylem area, which is usually the region closer to the cambium. The recommended distance by Burgess and Downey (2014) is 2.5 mm below the cambium but this might vary among species. An increment borer and a bark depth gauge can be used to assess bark thickness and xylem positioning.
• As sap flow is not constant across the radial sapwood profile (Miller et al 1980; Cohen et al 1981), it is recommended to place more thermometers in each probe to characterize the effect of sapwood depth in sap flow. The commercially available HRM probes have two thermistors in each probe, which allows for the use of the weighted average method (Hatton et al 1990) to integrate the sapflow across the xylem. This procedure is explained in details in the Integrating sap flow profile
• When drilling the holes, it is recommended to use a drill guide to assure that the holes are aligned and parallel. The symmetry of the holes can be assessed by inserting drill bits in the holes and inspecting them visually (Burgess and Downey 2014). Small misalignments might still be corrected using the procedure described in details in the Calculating Heat-Pulse Velocity

 

Heat-pulse energy and measurement time

• The HRM is based on the temperature ratio between probes (v₁/v₂); therefore, the absolute temperature of the heat pulse is not important. However, a minimum increase in temperature is necessary to allow the thermometers to register accurately temperature changes. On the other hand, high temperatures might excessively damage the tissues around the needle, thermally decoupling it from the sapwood and compromising the accuracy of the measurements. Burgess and Downey (2014) recommend keeping the maximum temperature rise after the heat-pulse in a range between 0.7° and 1.5° C.
• The changes in the v₁/v₂ should become approximately linear 60 s after the heat-pulse. Therefore, thermometer measurements should be conducted between 60 to 100 s after the heat-pulse. Multiple v1/v2 sampling is recommended to minimize random variation caused by thermal or electronic interference (Burgess et al 2001).

 

Calculating heat-pulse velocity

In HRM, the heat-pulse velocity (V_h; cm hr^-1) is calculated using the Marshal’s (1958) equation:

 

V_h= k/x ln⁡(v_1/v_2 )3600                                                                                      (1)

 

where k is the thermal diffusivity of fresh wood, x is the distance (in cm) between the heater and the probes, and v₁ and v₂ are the downstream and upstream temperature increases after the heat-pulse, that is:

 

v=T_h- T_0                                                                                                             (2)

 

where T_h is the probe temperature after the heat pulse and T₀ the probe temperature before the heat pulse.

 

Calculating thermal diffusivity

In the original HRM protocol, k is initially set to a nominal value of 2.5 × 10^–3 cm² s^–1 (Marshal 1958). Afterwards, Burgess et al (2001) proposed two independent methods to estimate actual k: simultaneously using CHPM and HRM to resolve for k (as CHPM does not require k it is possible to solve HRM and CHPM equations simultaneously). The results from this approach were similar to an analytical approach employed by Burgess et al (2001). In this approach, k is estimated using Marshall’s (1958) equation:

 

k=K_gw/(ρ_b c) 10000                                                                                            (3)

 

where K_gw is the green (fresh) sapwood thermal conductivity, ρ_b is sapwood density (kg m^-3; which can be calculated using this protocol) and c the specific heat capacity of fresh sapwood. The value of K_gw can be calculated using a modified Swanson (1983) equation:

 

K_gw= K_s m_c ρ_b/ρ_s +K_w (1-m_c ρ_b/ρ_s )                                                    (4)

 

where K_s is the water thermal conductivity (5.984 × 10^–1 J m^–1 s ^–1 °C^–1 at 20 °C), m_c is the sapwood water content (kg) and ρ_s is the density of water (999.97 kg m^-3). K_w is thermal conductivity of the dry wood matrix, calculated following Swanson (1983):

 

K_w=0.04182 (21-20(1-((ρ_b 0.6536+m_c)/1000)))                                               (5)

 

the value of c is estimated using a modified Edwards and Warwick (1984) equation:

 

c= (((w_d c_w + c_s ) (w_f-w_d))/w_f )                                                                   (6)

 

where c_w is the specific heat of the wood matrix (1200 J kg^-1 °C^-1 at 20°C), w_f is sapwood fresh weight and w_d is sapwood dry weight (kg).

Despite this approach producing results that agree well with the simultaneous solution and with the gravimetric validation experiments conducted by Burgess et al (2001), Vandegehuchte and Steppe (2012) suggest that the approach used by Burgess et al (2001) to calculate K_gw might produce errors of up to 10% in sap velocity estimates. According to Vandegehuchte and Steppe (2012) the mixture model used by Burgess can only be applied in sapwood with m_c lower than its fiber saturation point (i.e. the point where only bound water is present in the sapwood). As living, transpiring trees have a sapwood m_c higher than its fiber saturation point, Vandegehuchte and Steppe (2012) suggest the use of the following equation:

 

K_gw= K_w (m_c- m_{c_FSP} ) ) ρ_b/ρ_s +0.04186 (21-20 F_{v_FSP} )       (7)

 

where m_{c_FSP} is the sapwood water content at fiber saturation point and F_{v_FSP} is the void fraction at the fiber saturation point. The value of m_{c_FSP} can be determined following Roderick and Berry (2001):

 

m_{c_FSP} = 0.2 (ρ_b ρ_s ^-1)^-1/2                                                             (8)

 

while F_{v_FSP} is can be estimated with the following equation:

 

F_{v_FSP}=1 - G ((ρ_s/ρ_cw )+ m_{c_FSP})                               (9)

 

where G is sapwood specific gravity (ρ_b/ρ_s), and ρ_cw is cell wall density (1530 kg m⁻³, Kollmann and Côté (1968)).

 

Correction for probe misalignment

Probe misalignment in HRM is measured in situ, taking into account any thermal and physical asymmetries. In order to correct misalignment, V_h = 0 must be imposed. The most reliable way to impose V_h = 0 is to sever the stem or root with the probes to stop sap flow (Burgess et al 2001, Burgess & Dawson 2004, Oliveira et al 2005). Care must be taken to avoid moving the probes during the severing process; otherwise the correction would be made for the new sensor position and would be invalid. Some authors (Goldsmith et al 2013, Rosado et al 2012) assume that V_h during certain environmental conditions (usually very humid nights) approximates to zero, and use the data from these periods in the correction.

At V_h = 0, probe spacing is calculated as:

 

x_2= √(4kt ln(v_1/v_2 )+x_1^2 )                                                                           (10)

 

where x₂ is the incorrectly placed probe, x₁ is assumed to be correctly placed probe (at x cm from the heater) and t is the measurement time (i.e. the time interval in seconds after the heat pulse is emitted when the thermometer probes register the temperatures). Once x₁ and x₂ are estimated, corrected V_h is calculated as:

 

V_h= (4kt ln(v_1/v_2 )-(x_2^2 )+(x_1^2))/(2t(x_1-x_2)) 3600                               (11)

 

As it is usually unknown which probe is incorrectly positioned, it is recommended to solve equations 9 and 10 for x₁ and x₂, and average both solutions. This correction can be made only for a subset of the sap flow data. The corrected values should be linearly correlated with the uncorrected values; therefore, it is possible use the equation derived from a least-squares linear regression between corrected and uncorrected values to correct the remaining data (Burgess et al 2001).

 

Correction for wounding

The HRM probes can cause substantial mechanical damage in the wood. The probes interrupt the flow pathway where they were inserted, and the region surrounding the probes might become occluded by tylose formation (Barret et al 1995). This non-conducting wood area causes departure of the measured V_h. from actual V_h. This departure follows an approximately linear function; therefore, the wound corrected V_h (V_c) can be obtained simply by multiplying V_h by a correction coefficient (B) based on wound width (Burgess et al 2001):

V_c=V_h B                                                                                                            (12)

wound width can be determined by following this protocol. The B for commonly found wound widths for 0.5 mm probe configurations is given in the table below (reproduced from Burgess et al 2001):

 

Wound width (cm)	Correction coefficient (B)
0.17	1.7283
0.18	1.7853
0.19	1.8568
0.20	1.9216
0.21	1.9891
0.22	2.0594
0.23	2.1326
0.24	2.1825
0.26	2.3176
0.28	2.4813
0.30	2.6383

 

 

Correction for stem temperature variation

Recently, Vandegehuchte et al (2014) have demonstrated that temporal variations in stem temperature might create errors of up to 40% in HRM V_h. These errors are attributed to the time interval (60-100 s) between the heat-pulse and the thermometer measurements. Vandegehuchte et al (2014) propose a simple solution to these errors by including a temperature correction in the data obtained by HRM.

To do that, Vandegehuchte et al (2014) determined the slope and intercept of the stem temperature variations over a 30 s period before the heat-pulse. The projected values of temperature  were subtracted from the measured temperature values during and after the heat-pulse application. Therefore, v in equation 2 would become:

 

v =(T_h - ̂ T_h)-(T_0 - ̂ T_0 )                                   (13)

 

Calculating sap velocity

The V_c is composed by the weighted average of the velocity of moving sap and the stationary wood in the xylem (Marshall 1958). Therefore, we can relate V_c to actual sap velocity (V_s) by measuring the fractions of sap and wood in the xylem (basically, the sapwood water content), and accounting for their different densities and specific heat. Burgess et al (2001) uses a modified Marshall’s (1958) equation to calculate V_s (cm hr^-1):

 

V_s= (V_c ρ_b (c_w+m_c c_s))/(ρ_s c_s )                                                               (14)

 

where c_s is water specific heat (4182 J Kg^-1 °C^-1 at 20°C).

 

Calculating sap flow

Sap flow (Q; cm³ hr^-1) is easily derived from V_s by multiplying it by the total xylem cross-sectional area (CSA; cm²):

 

Q=V_s CSA                                                                                                           (15)

 

estimating CSA can be done by following one of the methods described here. Alternatively, when it is hard to differentiate between active and inactive xylem, it is possible to use the method used by Goldstein et al (1998). In this method, dye is injected in holes made with an increment borer (other authors prefer to use drills or a chisel and mallet) at the base of the tree. After some time (1-2 h), a sample is collected with an increment borer 2-4 cm above the point where the dye was injected. The area of conducting tissue is determined by the extension of the sapwood colored by the dye.

 

Integrating sap flow depth profile

Assuming that sap flow is uniform along the sapwood depth profile might produce large errors (-90 to 300%; Nadezhdina et al 2002). One common and robust method for integrating sap flow depth profile is to use the weighted average method (Hatton 1990). In this method, if a set of n sensors are placed at different depths between the cambium and the heartwood, the tree total CSA is divided into n concentric annuli. The inner radius (r_k) of a given annulus k, occurs midway between sensors k and k+1, where k = 1 ,…, n-1. This way, all points within the annulus k are closer to the sensor k than any other sensor; then we assume that V_s captured by sensor k (V_sk) represents the entire annulus k. Following this logic, the total sap flow of a tree being monitored by n sensors is the sum of the sap flow of its n concentric annuli:

 

Q= ∑_k^n₌₁ π (r_k^2-r_k^2_-1)V_sk                           (16)

 

Hatton (1990) recommends keeping the area of each annuli constant, so every sensor contributes with an equal amount of information to Q profile. Therefore, the spacing between sensors should decrease, as they get closer to the cambium. More details on how to space the sensors to achieve this can be found on Hatton (1990).

Links to resources and suppliers

www.ictinternational.com

Literature references

Barrett DJ, Hatton TJ, Ash JE, Ball MC (1995) Evaluation of the heat pulse velocity technique for measurement of sap flow in rainforest and eucalypt forest species of south-eastern Australia. Plant Cell Environment 18, 463–469.

Becker P (1996) Sap flow in Bornean heath and dipterocarp forest trees during wet and dry periods. Tree Physiology 16, 295–299.

Becker P (1998) Limitations of a compensation heat pulse velocity system at low sap flow: implications for measurements at night and in shaded trees. Tree Physiology 18, 177–184.

Burgess SSO, Adams MA, Turner NC and Ong CK (1998) The redistribution of soil water by tree root systems. Oecologia 115, 306–311.

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, Bleby TM (2006) Redistribution of soil water by lateral roots mediated by stem tissues. Journal of Experimental Botany 57, 3283-3291.

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 & Environment 27, 1023-1034.

Burgess SSO, Downey A (2014) SFM1 Sap flow meter manual Version 4.0. ICT International Pty. Ltd. http://www.ictinternational.com/content/uploads/2014/03/SFM1-complete-manual1.pdf.

Cohen Y, Fuchs M and Green GC (1981) Improvement of the heat pulse method for determining sap flow in trees. Plant, Cell & Environment 4, 391-397.

Dawson TE, Burgess SSO, Tu KP, Oliveira RS et al. (2007). Nighttime transpiration in woody plants from contrasting ecosystems. Tree Physiology 27, 561-575.

Edwards WRN, Becker P, Cermak J (1996) A unified nomenclature for sapflow measurements. Tree Physiology 17, 65-67.

Edwards WRN, Warwick NWM (1984) Transpiration from a kiwifruit vine as estimated by the heat pulse technique and the Penman-Monteith equation. New Zealand Journal of Agricultural Research 27, 537–543.

Eller CB, Lima AL, Oliveira, RS (2013) Foliar uptake of fog water and transport belowground alleviates drought effects in the cloud forest tree species, Drimys brasiliensis (Winteraceae). New Phytologist 199, 151-162.

Goldsmith GR, Matzke NJ, Dawson TE (2013) The incidence and implications of clouds for cloud forest plant water relations. Ecology letters 16, 307-314.

Goldstein G, Andrade JL, Meinzer FC, Holbrook NM, Cavelier J, Jackson P and Celis A (1998) Stem water storage and diurnal patterns of water use in tropical forest canopy trees. Plant Cell Environment 21, 397–406

Green SR, Clothier BE (1988) Water use of kiwifruit vines and apple trees by the heat-pulse technique. Journal of Experimental Botany 39, 115–123.

Hatton TJ, Catchpole EA, Vertessy RA (1990) Integration of sapflow velocity to estimate plant water use. Tree Physiology 6, 201-209.

Kollmann FFP, Côté WA (1968) Principles of wood science and technology 1, Solid wood. Springer, New York, pp 246–250.

Lott JE, Khan AAH, Ong CK and Black CR (1996) Sap flow measurements of lateral tree roots in agroforestry systems. Tree Physiology 16, 995–1001.

Marshall DC (1958) Measurement of sapflow in conifers by heat transport. Plant Physiology 33, 385-396.

Miller DR, Vavrina CA, Christensen TW (1980) Measurement of sap flow and transpiration in ring-porous oaks using a heat pulse velocity technique. Forest Science 26, 485-494.

Nadezhdina N, Čermák, J, Ceulemans R (2002) Radial patterns of sap flow in woody stems of dominant and understory species: scaling errors associated with positioning of sensors. Tree Physiology 22, 907-918.

Oliveira RS, Dawson TE, Burgess SS, Nepstad, DC (2005) Hydraulic redistribution in three Amazonian trees. Oecologia 145, 354-363.

Roderick ML, Berry SL (2001) Linking wood density with tree growth and environment: a theoretical analysis based on the motion of water. New Phytologist 149, 473–485.,

Rosado BH, Oliveira RS, Joly CA. Aidar MP, Burgess SSO (2012) Diversity in nighttime transpiration behavior of woody species of the Atlantic Rain Forest, Brazil. Agricultural and Forest Meteorology 158, 13-20.

Swanson RH (1983) Numerical and experimental analyses of implanted-probe heat pulse velocity theory. Ph.D. Thesis, University of Alberta, Edmonton, Canada, 298 p

Swanson RH, Whitfield DWA (1981) A numerical analysis of heat pulse velocity theory and practice. Journal of Experimental Botany 32, 221–239.

Vandegehuchte MW, Burgess SS, Downey A, Steppe K (2014) Influence of stem temperature changes on heat pulse sap flux density measurements. Tree Physiology, doi: 10.1093/treephys/tpu068.

Vandegehuchte MW, Steppe K (2012) Improving sap flux density measurements by correctly determining thermal diffusivity, differentiating between bound and unbound water. Tree Physiology 32, 930-942.

Zotz G, Patino S and Tyree MT (1997) Water relations and hydraulic architecture of woody hemiepiphytes. Journal of Experimental Botany 48, 1825–1833.