Using electrical resistivity tomography to differentiate sapwood from heartwood: application to conifers

• Published in: Tree physiology Vol. 33; no. 2; pp. 187 - 194
• Main Authors: Guyot, Adrien, Ostergaard, Kasper T, Lenkopane, Mothei, Fan, Junliang, Lockington, David A
• Format: Journal Article
• Language: English
• Published: Canada 01-02-2013
• Subjects: Algorithms; Electric Impedance; Electrodes; Electrolytes - analysis; Linear Models; Pinus - chemistry; Pinus - physiology; Plant Transpiration - physiology; Queensland; Tomography - instrumentation; Tomography - methods; Trees; Water - physiology; Wood - chemistry; Wood - physiology; Xylem - chemistry; Xylem - physiology; Queensland
• ISSN: 0829-318X; 1758-4469
• DOI: 10.1093/treephys/tps128

Estimating sapwood area is one of the main sources of error when upscaling point scale sap flow measurements to whole-tree water use. In this study, the potential use of electrical resistivity tomography (ERT) to determine the sapwood-heartwood (SW-HW) boundary is investigated for Pinus elliottii Engelm var. elliottii × Pinus caribaea Morelet var. hondurensis growing in a subtropical climate. Specifically, this study investigates: (i) how electrical resistivity is correlated to either wood moisture content, or electrolyte concentration, or both, and (ii) how the SW-HW boundary is defined in terms of electrical resistivity. Tree cross-sections at breast height are analysed using ERT before being felled and the cross-section surface sampled for analysis of major electrolyte concentrations, wood moisture content and density. Electrical resistivity tomography results show patterns with high resistivities occurring in the inner part of the cross-section, with much lower values towards the outside. The high-resistivity areas were generally smaller than the low-resistivity areas. A comparison between ERT and actual SW area measured after felling shows a slope of the linear regression close to unity (=0.96) with a large spread of values (R(2) = 0.56) mostly due to uncertainties in ERT. Electrolyte concentrations along sampled radial transects (cardinal directions) generally showed no trend from the centre of the tree to the bark. Wood moisture content and density show comparable trends that could explain the resistivity patterns. While this study indicates the potential for application of ERT for estimating SW area, it shows that there remains a need for refinement in locating the SW-HW boundary (e.g., by improvement of the inversion method, or perhaps electrode density) in order to increase the robustness of the method.