TDP is a constantly heated sap flow principle that assumes there is no background temperature gradient between the probes and the plant (thermal insulation is required around the sensors and the stem). However, the data shows that Natural Thermal Gradients (NTG) from the soil (which cannot be incorporated into the TDP measurement) are significant (1°C or greater) hence, the temperature gradient measured is not a true indication of the actual sap flow. The result is a substantial overestimation due to the background temperature gradient in the mornings and again underestimation in the evenings, none of which are of a constant magnitude to even begin to attempt compensating for.
Another cause of underestimation of sap velocity using TDP is the thickness of the sapwood. The standard TDP-30 sensor has a length of 30 mm consisting of a full length line heater and only a single thermocouple located at the halfway point (15 mm). The sensor design is adequate for use in boreal Pinus species with thick sapwood of at least 30 to 40 mm, but is inadequate for most other species. TDP sensors cannot be used to accurately measure sap flow of Eucalyptus species for example, because on average the sapwood thickness is approximately 20 to 25 mm thick. Therefore, the TDP sensor will almost always have the heater extend significant depths into heartwood or non-conducting xylem. This results in highly variable and underestimated sap velocity because the non-conducting xylem (heartwood) artificially increases the dTmax as it does not dissipate the heat.
With these two conflicting variables causing both positive and negative errors of unknown quantities, it is clear that the TDP principle has significant limitations in application; an opinion that is supported by the majority of papers published using the TDP technique. Papers published from the 7th International Sap Flow Workshop, Seville Spain 2008, that have grappled with these issues rather than successfully using the TDP sensor as a tool for researching plant water relations include: Sevanto, et.al., Conceicao & Ferreira, Ferreira et.al., and Chavarro, et.al.
Alternatively, the Heat Ratio Method (HRM) sensor is a modified heat pulse technique that consists of three needles: two measurement needles located equidistant above and below a central line heater. Because it is a pulsed technique using a short 2 to 6 second pulse of heat and a 100 second measurement window, the effect of NTGs are avoided as ambient temperature changes within the measurement interval are insignificant or non-existent. This has additional benefits in practical deployment as the sensors also require no thermal shielding to thermally isolate the measurement site against NTGs as with TDP.
The HRM needles are 35 mm in length and have two temperature measurement locations, 7.5 mm from the tip and 22.5 mm from the tip. This provides a measurement of radial velocity across the sapwood and quantifiably measures the velocity gradient and/or identifies when the needle extends into non-conducting xylem (so the inner measurement point can be discounted).
The HRM can measure very low flow (approx 1 cm hr-1), zero flow, and reverse flow in a range of stem sizes, making it very robust and flexible. The TDP (as with most other commercial sap flow methods) cannot measure zero or reverse flow and typically only measures to a minimum level of ~4 cm hr-1 at low flow – which on average results in approximately 40% of all sap flow not being recorded.
In a world of ever decreasing water availability, the need to accurately quantify plant water relations and screen for plants with the ability to redistribute water within their growing environment through mechanisms such as hydraulic lift and reverse flow, is crucial. The requirement to adopt new techniques and methodologies that can measure these mechanisms is paramount.
|Measurement Units||Heat Pulse Velocity (cm hr-1)
Sap Velocity (cm hr-1)
Sap Flow (g hr-1)
|Sap Flux Density (cm3 hr-1 cm-2)|
|Measurement Range||-20 to 60 cm hr-1||0 to 80 cm3 cm-2 hr-1 (with calibration)|
|Measurement Accuracy||0.5 cm hr-1||Not specified*|
|Measurement Resolution||0.01 cm hr-1||1 µV|
|Measures Reverse Flow||Yes||No|
|Measures Low Flow||Yes||Standard minimum measurable velocity is approx. 4 cm hr-1, can measure to 0 cm hr-1 with empirical calibration|
|Measures High Flow||Yes||Yes|
|Measures Multiple Radial Points||Yes – 2 independent measurement points in the same radial profile, spaced 15 mm apart. Can be used to characterise flow in the inner and outer xylem.||No – Standard sensor is one measurement point|
|Used on large diameter stems||Yes – For trees of any diameter, but only for the outer 4 0mm of xylem. Using 2 measurement points, flow can be characterised in the inner and outer xylem, enabling correction for radial variation across sapwood.||Yes – Only for the outer 20 mm of xylem|
|Used on Small Diameter Stems||Yes – Stems larger than 10mm||No – Sensor needle must be fully inserted and insulated to prevent errors due to Natural Thermal Gradients|
|Used on Roots||Yes – Has been used to study hydraulic lift and hydraulic redistribution in root systems||No – Does not measure reverse flow, unsuitable for installation in soil|
|Is wound response accounted for?||Yes||No|
|Is wound response relevant?||Yes – All species occlude cells and repair intrusive wounds, which causes a non-conductive zone of tissue that effects heat transfer and ultimately measurement sensitivity and accuracy if not corrected for while processing raw data.|
|Heat Source||Heat Pulse||Continuous|
|Requires radiation shielding?||No – Measurement time is so short that no significant change in air temperature or effect of direct incident radiation will occur that can effect the measurement.||Yes – Requires thermal and radiation shielding around the installation site, and 50 cm to 1 m both above and below the installation, preferably to ground level.|
|Effected by Natural Thermal Gradients (NTG)?||No||Yes – Causes error in measurement accuracy and sensitivity|
|Does the sensor need to be inserted only in sapwood?||No – multiple measurement points enable measurement of radial gradient and determination of sapwood/heartwood border should the needles extend beyond sapwood into heartwood.||Yes – crucial that the entire length of the needle be in sapwood. If the needle extends into heartwood or non-conducting xylem heat will not be dissipated and errors of up to 50% can occur.|
|Do heat lags effect the measurement?||No – A short heat pulse is generated for each measurement, eliminating the effect of ambient thermal conditions.||Yes – the sensor is continuously heated|
|Data Processing & Analysis|
|Calibration method||Specific wood properties and wound coefficients to convert heat pulse velocity to corrected sap velocity and sap flow.||Empirical calibration requiring data corrections for dTmax; using extensive data modelling.|
|Raw Data Processing Required||No – Units measured are cm hr-1||Yes – Conversion from analogue uV to temperature, then conversion to sap velocity|
|Software||Yes – Sap Flow Tool and ICT Combined Instrument Software. Data files are in CSV format, and can be used in Excel||No – Excel only, requires processing.|
|Data Output||Raw Heat Pulse Velocity (cm hr-1)
Corrected Sap Velocity (cm hr-1)
Corrected Sap Flow (g hr-1)
|Temperature Difference (uV/mV or °C)|
|Temporal Logging Resolution||Minimum 3 minutes to ensure heat dissipation, standard 15 minutes.||Minimum 1 minute, typically 15 to 60 minutes to prevent measurement noise|
|Number of sensors per logging system||N/A (standalone)||32 (with multiplexer)|
|Communications||USB, 2.4GHz wireless.||RS-232 Serial|
|Memory Capacity||MicroSD: 4GB standard (>100 years), expandable up to 16GB||4MB (1 year hourly data; sap flow calculations: 8 months for 8 gages, 2 months for 32.), expandable up to 8MB|
* Underestimations of actual flux density, ranging between 6 and 90%, have been reported for a wide variety of species when comparing TDP to other methods (Lundblad et al. 2001; Bovard et al. 2005; Silva et al. 2008; Iida and Tanaka 2010) or during new calibration experiments on excised stem or branch segments (de Oliveira Reis et al. 2006; Taneda and Sperry 2008; Bush et al. 2010; Hultine et al. 2010; Steppe et al. 2010), cut trees (Lu and Chacko 1998; Uddling et al. 2009) or potted plants (Braun and Schmid 1999; McCulloh et al. 2007).