Novel, cyclic heat dissipation method for the correction of natural temperature gradients in sap flow measurements. Part 2. Laboratory validation

Abstract

Sap flow measurements conducted with thermal dissipation probes (TDPs) are vulnerable to natural temperature gradient (NTG) bias. Few studies, however, attempted to explain the dynamics underlying the NTG formation and its influence on the sensors' signal. This study focused on understanding how the TDP signals are affected by negative and positive temperature influences from NTG and tested the novel cyclic heat dissipation (CHD) method to filter out the NTG bias. A series of three experiments were performed in which gravity-driven water flow was enforced on freshly cut stem segments of Fagus sylvatica L., while an artificial temperature gradient (ATG) was induced. The first experiment sought to confirm the incidence of the ATG on sensors. The second experiment established the mis-estimations caused by the biasing effect of the ATG on standard TDP measurements. The third experiment tested the accuracy of the CHD method to account for the ATG biasing effect, as compared with other cyclic correction methods. During experiments, sap flow measured by TDP was assessed against gravimetric measurements. The results show that negative and positive ATGs were comparable in pattern but substantially larger than field NTGs. Second, the ATG bias caused an overestimation of the standard TDP sap flux density of ∼17 cm(3) cm(-2) h(-1) by 76%, and the sap flux density of ∼2 cm(3) cm(-2) h(-1) by over 800%. Finally, the proposed CHD method successfully reduced the max. ATG bias to 25% at ∼11 cm(3) cm(-2) h(-1) and to 40% at ∼1 cm(3) cm(-2) h(-1). We concluded that: (i) the TDP method is susceptible to NTG especially at low flows; (ii) the CHD method successfully corrected the TDP signal and resulted in generally more accurate sap flux density estimates (mean absolute percentage error ranging between 11 and 21%) than standard constant power TDP method and other cyclic power methods; and (iii) the ATG enforcing system is a suitable way of re-creating NTG for future tests.

Optimization of dry‐season sap flow measurements in an oak semi‐arid open woodland in Spain

“…It followed a strict order: (1) NTG bias correction, (2) night flow Δ T max correction and (3) radial‐azimuthal correction. In step 1, the standard TDP measurements (TDP std ) were obtained first, and then, the cyclic heat dissipation (CHD) correction method (Lubczynski et al ., ; Reyes‐Acosta et al ., ) was applied (TDP NTG‐free ). In step 2, the TDP NTG‐free data were corrected for night flow Δ T max (TDP Δ T max‐corr ) with help of an HFD reference (HFD 2cm ).…”

“…(). Hence, the TDP method is particularly sensitive to NTG if applied in sparsely vegetated woodlands or savannahs and when exposed to highly contrasting day/night temperatures (Lubczynski et al ., ; Reyes‐Acosta et al ., ), a common situation in the Sardón study area in dry seasons.…”

“…However, there is a general concern about the efficiency and versatility of these methods when used to quantify the actual amount of water used by plants or in higher levels of organization such as tree stands or catchments. The thermal methods are prone to inaccuracies if certain aspects of tree dynamics are not accounted for, such as natural temperature gradients (NTGs) (Do and Rocheteau, ; Lu et al ., ; Lubczynski et al ., ; Reyes‐Acosta et al ., ), night flow (Grime and Sinclair, ; Regalado et al ., ), azimuthal/radial variability of the sap flow (Čermák et al ., ; Lu et al ., ) and reverse flow (Burgess et al ., ). Each inaccuracy can be accounted for with an appropriate method, but there is not a single universal thermal method that can handle all the inaccuracies.…”

“…The TDP method has five potential inaccuracies that should be taken into account while quantifying J p : (1) in standard implementation, the TDP method is vulnerable to NTG bias resulting in data that reflect changes related not only to the target heat convection of flowing water but also to the measurement of NTG-related heat-conduction bias ( Cermák and Ku cera, 1981;Rocheteau, 2002a, 2002b;Lu et al, 2004;Lubczynski et al, 2012;Reyes-Acosta et al, 2012); (2) when using TDP, it is necessary to define zero flow condition, corresponding with the maximum measured temperature difference (ΔT max ) usually observed at night or pre-dawn; however in trees with active night flow, assigning ΔT max without correction leads to an incorrect estimation of sap flow (Lu et al, 2004;Regalado and Ritter, 2007); (3) TDP measurements cannot provide radial and azimuthal variability of J p (Lu et al, 2004) in trees with sapwood depth >2 cm; (4) standard TDP does not measure bi-directional flows, e.g. hydraulic redistribution (Burgess et al, 2000); (5) the TDP method uses an empirical calibration (Granier, 1985) relating the measured ΔT to J p that is appropriate for diffuse-porous sapwood tree species, but according to Bush et al (2010) ring-porous, tree species might not follow such calibration (Reis et al, 2006;Steppe et al, 2010).…”

Comment on “Sap flow measurements by a single thermal dissipation probe: exploring the transient regime” — Ann. For. Sci. 66 (2009) by Mahjoub et al. and “Sap flow measurement by a single thermal dissipation probe in transient regime: implementation of the method and test under field conditions” — Ann. For. Sci. 1-9 (2012) by Masmoudi et al.

“…Recently, large improvements have been made by taking into account natural temperature gradients and assessing new calibration factors (Do and Rocheteau 2002a;Do and Rocheteau 2002b;Isarangkool Na Ayutthaya et al 2009;Reyes-Acosta et al 2012;Steppe et al 2010). Besides, Do et al (2011) suggested to further optimize the TDP design by applying only a single needle with a cyclic heating scheme.…”