Quantification of uncertainties introduced by data-processing procedures of sap flow measurements using the cut-tree method on a large mature tree
Introduction
Heat-based sap flow methods have become a principal means to non-destructively provide whole-tree water use estimates across spatiotemporal scales (Hassler et al., 2018), while also improving our understanding of the impact(s) of a changing climate (Berdanier and Clark, 2018; Zhang et al., 2016) and forest management practices (Doody et al., 2015) on forested landscapes. Several sap flow methods are available to measure tree-water use (Fernández et al., 2017) using heat as a tracer for sap flow. Calculating whole-tree sap flux rates using heat tracers requires several steps of data-processing (Looker et al., 2016; Peters et al., 2018). Proper estimation of the sapwood thermal diffusivity as well as corrections for probe misalignment and wounding need to be considered (Burgess et al., 2001; Burgess and Downey, 2018; Marshall, 1958). Zero-flow determination is an additional crucial step, for which a range of empirical, systematic (Granier, 1987; Lu et al., 2004), and physically-based (Burgess and Downey, 2018; Oishi et al., 2008) methods are available. Additionally, strong radial and circumferential variations (Van de Wal et al., 2015) are often ignored, potentially leading to over- or under-estimating whole-tree water use when extrapolating a single sensor measurement across the entire stem. A global assessment of the impact of the combination of those data-processing steps on sap flow studies is currently lacking (but see Looker et al., 2016; Peters et al., 2018; Rabbel et al., 2016; Van de Wal et al., 2015 among others, addressing individual data-processing approaches), challenging the interpretation, comparison, and integration of individual tree-based studies when used in large-scale ecohydrology models.
A general quantification of the uncertainties generated by both the sap flow sensor method itself and the combination of methodological decisions and data-processing procedures on large mature trees is lacking, but crucial (Flo et al., 2019), particularly when attempting to derive landscape-scale processes (e.g. ecosystem water and carbon fluxes, catchment hydrological and vegetation responses to climate and land-use changes; Bonan, 2008; Schlesinger and Jasechko, 2014). Calibration studies addressing the accuracy of the sap flow measurements in woody plants have been previously conducted (see the Supplementary Material in Flo et al. (2019) for a list of studies and calibration materials), however, many of them were mostly focused on smaller trees (less than 10 cm in diameter), shrubs, or branch segments, which are unlikely to provide reliable or appropriate estimates of water-use in larger diameter trees. To date, only a few studies have directly assessed the accuracy of sap flow sensors in estimating the water uptake of large mature trees (Olbrich, 1991; Roberts, 1977; Vertessy et al., 1997). However, the height (>10 m), diameter (>20 cm) and weight of mature forest trees presents experimental and logistical challenges for obtaining a direct measurement of water uptake to compare with the sap flow sensors’ estimates, and thus, these studies are very limited in scope. The anatomical, morphological, and physiological variability in the stem of large mature trees can result in potentially strong diverging patterns between and within-trees in sap flux densities (Cohen et al., 2008; Shinohara et al., 2013; Van de Wal et al., 2015). The choice of the method in correcting for zero-flows and extrapolating to whole-tree sap flux rates may increase the error in sensor-based estimates of water use of an individual tree. Ultimately, this can jeopardize the interpretation of climate-response analyses and stand-level water uptake estimates for large and mature trees across temporal and spatial scales.
In this study, we aimed to: (1) quantify the extent of uncertainty in sap flow measurements related to circumferential variability, the application of different sapwood area estimation methods and zero-flow corrections; and (2) assess the impact of each and the combination of these factors on the relationship with climate variables and estimates of cumulative whole-tree water uptake. We used heat ratio method (HRM) sap flow heat-pulse sensors (Burgess et al., 2001) which have shown relatively good performance in accuracy when compared to other sap flow methods (Flo et al., 2019). Furthermore, we used the cut-tree technique to directly measure trunk water uptake gravimetrically by immersing the stem of a large tree (20 m tall, 35 cm in diameter) in water (Roberts, 1977). Using this technique, we can directly compare the HRM sap flux estimates to the gravimetrically-measured water uptake, and then quantitatively assess the uncertainty and error associated with several data-processing methods applied to the sap flow data.
Section snippets
Study site and climate data
The study was carried out in Thorhild County (54°17′11.6″ N, 112°46′08.6″ W), approximately 120 km northeast of Edmonton, Alberta, Canada. The site is located within the dry mixedwood natural sub-region of the boreal region, composed of a mixture of trembling aspen (Populus tremuloides Michx.) and white spruce (Picea glauca (Moench) Voss), surrounded by agricultural lands. The selected trees for the study were part of a 60 - 70-year-old trembling aspen dominated stand (~ 600 stems per hectare).
Sap flow response to the cutting
Immediately after the cut and for the next 48 h, sap flux density (Js) was similar to measurements taken before the cut and to those measured in the three control trees (Fig. 1a). Sap flux density then declined in the focus tree over the next eight days compared to the three control trees, which all maintained higher Js. Interestingly the crown of our focus tree showed no sign of hydraulic damage for the duration of the study (i.e. wilting or color changes; Fig. 1b and c). Transpiration rate of
Discussion
We quantified the uncertainties introduced by the placement of sap flow sensors (cardinal directions) and the application of different data-processing approaches (radial integration method used, zero-flow correction, and sapwood area estimation methods) on sap flow estimates for large diameter trees. These estimates were compared to gravitational measurements using the cut-tree method. The temporal precision of sensor-derived fluxes with the gravimetrically-measured water uptake was high when
Conclusion
This cut-tree study of sap flux measurements on a large mature tree using the HRM sap flow sensors highlighted that uncertainties arise from several methodological steps from sensor installation to the calculation of hourly sap fluxes. Although the cut-tree method provides a unique way to directly obtain gravimetric measurement of water uptake at a fine temporal scale for large trees (Olbrich, 1991; Roberts, 1977; Smith, 1992; Vertessy et al., 1997), these studies are complex to set up and
CRediT authorship contribution statement
Morgane Merlin: Conceptualization, Methodology, Investigation, Formal analysis, Visualization, Writing - original draft. Kevin A. Solarik: Conceptualization, Methodology, Investigation, Writing - review & editing. Simon M. Landhäusser: Funding acquisition, Project administration, Supervision, Conceptualization, Methodology, Investigation, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
We thank Ian Jamieson for his invaluable help and expertise during planning and construction of the wooden structure, as well as Peter Lazowski for his help in the field. We thank all those who provided field and logistical support for this project (Frances Leishman, Robert Hetmanski, Natalie Scott, Ashley Hart, and Julie Zettl) as well as laboratory work (Pak Chow) and data-processing support (Newton Tran, ICT International Pty Ltd.). We thank Cinnamin Landhäusser for her support and
References (51)
- et al.
A synthesis of bias and uncertainty in sap flow methods
Agric. For. Meteorol.
(2019) - et al.
Calibration and comparison of thermal dissipation, heat ratio and heat field deformation sap flow probes for diffuse-porous trees
Agric. For. Meteorol.
(2017) - et al.
Contribution of sapwood traits to uncertainty in conifer sap flow as estimated with the heat-ratio method
Agric. For. Meteorol.
(2016) - et al.
Baseliner: an open-source, interactive tool for processing sap flux data from thermal dissipation probes
SoftwareX
(2016) - et al.
Estimating components of forest evapotranspiration: a footprint approach for scaling sap flux measurements
Agric. For. Meteorol.
(2008) - et al.
Transpiration in the global water cycle
Agric. For. Meteorol.
(2014) The heat pulse velocity technique for determining water uptake of Populus deltoides
South African J. Bot.
(1992)- Alberta Climate Information Service (ACIS), 2018. Interpolated weather data since 1961 for Alberta townships [WWW...
- et al.
Tree water balance drives temperate forest responses to drought
Ecology
(2018) - et al.
Predictive models for radial sap flux variation in coniferous, diffuse-porous and ring-porous temperate trees
Tree Physiol.
(2016)
A validation, comparison and error analysis of two heat-pulse methods for measuring sap flow in Eucalyptus marginata saplings
Funct. Plant Biol.
Forests and climate change: forcings, feedbacks, and the climate benefits of forests
Science
An improved heat pulse method to measure low and reverse rates of sap flow in woody plants
Tree Physiol.
SFM1 Sap Flow Meter Manual
Sap flow measurements with some thermodynamic methods, flow integration within trees and scaling up from sample trees to entire forest stands
Trees
Seasonal changes of azimuthal, radial, and tree-to-tree variations in sap flux affect stand transpiration estimates in a Cryptomeria japonica forest, central Taiwan
J. For. Res
Variations in the radial gradient of sap velocity in trunks of forest and fruit trees
Plant Soil
Quantifying water requirements of riparian river red gum (Eucalyptus camaldulensis) in the Murray-Darling basin, Australia - implications for the management of environmental flows
Ecohydrology
How reliable are heat pulse velocity methods for estimating tree transpiration?
Forests
Evaluation of transpiration in a Douglas-Fir stand by means of sap flow measurements
Tree Physiol.
Tree-, stand- and site-specific controls on landscape-scale patterns of transpiration
Hydrol. Earth Syst. Sci.
Integration of sapflow velocity to estimate plant water use
Tree Physiol.
Scaling from single-point sap velocity measurements to stand transpiration in a multispecies deciduous forest: uncertainty sources, stand structure effect, and future scenarios
Can. J. For. Res.
Cited by (17)
Differentiated responses of daytime and nighttime sap flow to soil water deficit in a larch plantation in Northwest China
2023, Agricultural Water ManagementCharacteristics of soil evaporation at two stages of growth in apple orchards with different ages in a semi-humid region
2023, Agricultural Water ManagementIncreases in sap flow and storage can compensate for successively greater losses of conducting area in large trees
2023, Agricultural and Forest MeteorologyComparisons of three scaling up methods to estimate stand transpiration of a xerophytic shrub (Salix psammophila) in northern China
2022, Journal of HydrologyCitation Excerpt :In the S. psammophila stand, estimated daily mean stand T based on leaf area, cross-sectional area and number of branches was always higher than the actual value (p < 0.05), suggesting consistent overestimation. This result is not consistent with the general perception that SF measurements gave stand T values similar to those obtained by other T measurements (Granier, 1987; Yue et al., 2008; Merlin et al., 2020). In this study, the consistent overestimation between actual and estimated stand T based on different methods might be due to that SF measurements largely suffered from different potential sources of error related to intrinsic limitation of the SF methods and/or to how these methods are applied (Mitchell et al., 2009; Steppe et al., 2015; Vandegehuchte and Steppe, 2013).