Electric resistance tomography and stress wave tomography for decay detection in trees—a comparison study

Field testing

The study was conducted at the Northeast Forestry University Experimental Forest Farm, Harbin, Heilongjiang Province, China. The area is located at longitude 126°37′E, latitude 45°43′N, 140 m above sea level, and a slope of 5°. The study area is 43.95 hectares, located in the warm temperate zone semi-humid monsoon climate zone; the annual average temperature is 3.6 °C. The highest temperature in July was 36.4 °C and the lowest temperature in January was −38.1 °C. In the frost-free period, the average rainfall in the area is 600 mm/year. By reforestation between the early 1950s and the end of the 1960s, trees mixed species have been divided into 46 sample plots, each with different type of trees.

In July 2017, in the experimental forest, an experienced forester visually identified 30 Manchurian ash and 30 Populus simonii standing trees that could potentially have internal decay, then 10 solid Manchurian ash and 10 solid Populus trees were also selected, they all were marked in order (Number 1 to number 30 were decay, and number 31 to 40 were solid). All selected trees were between the ages of 50∼60 years old. The DBH (diameter at breast height) of Manchurian ash trees was 20∼38cm, and that of Populus trees was 30∼50 cm.

Nondestructive testing of trees

The instruments used were: PICUS Tree Tronic Electrical Resistance Tomography (ERT) (Argus GmbH, Germany), Arbotom Stress Wave Tomography (SWT) (RINNTECH GmbH, Germany).

ERT measurements were conducted at 100 cm above the ground. The Picus ERT measurement system consisted of 12 electrodes that were evenly placed around the trunk along the horizontal plane during the test. For ease of analysis, the first electrode was arranged in south and the other electrodes were arranged equidistantly in a clockwise manner. Each electrode was clipped and attached to a nail (2 mm in diameter) that had been tightly inserted into the bark and sapwood. After the electrodes were connected, resistance tests were started. In order to reduce the error, two tests were performed at each tree, and the current histograms displayed on the instrument were observed. If there were two differences, a third test must be performed, and then the output file name was recorded for subsequent analysis. Upon completion of ERT measurements, 2D tomograms were obtained combined with the Picus Q72 software program.

After ERT inspections were completed, SWT measurements were carried out (the same positions as for ERT measurement). All sensors were located almost the same as ERT sensors’ positions in the trees, and the transducers were connected at an angle of 90° to the trunk’s longitudinal axis to detect the propagating travel time. Measurements were repeated with the transmitter probe at each point. A complete data matrix was obtained using this measurement process at each test location. These measurements are used as input to the system software. Due to differences in species and paths, two-dimensional (2D) tomographic images were acquired from the original stress wave transmission time using ARBOTOM software to understand the experimental values in this study.

Wood cores obtaining

Wood cores were extracted from the two directions of each cross section using a Swedish wood core drill (diameter 6 mm): one was from the south to the north in the radial direction and the other was from the east to the west in the radial direction, as shown in Fig. 1. When a wood core was decayed, we would get a solid core nearby for comparison. The wood core samples were shown in Fig. 2. After being extracted, the wood cores were immediately put in ziplock bags and were taken back to the laboratory. They were divided into pieces of 1 cm sections, and then they were dried to constant weight at 70 °C in an electric blast oven and weighed.

Calculations and data analysis

Calculation of the severity of decay detected by ERT and SWT

Based upon nondestructive testing of ERT and SWT, the electrical resistance 2D tomograms and stress wave velocity 2D tomograms were acquired. The healthy trees of the same species have the similar 2D tomograms, and decay trees have different 2D tomograms.

In order to quantitatively evaluate the ER tomograms of sample trees, all corresponding electrical resistances (ERs) at each pixel in the tomogram were further calculated by visualization and inversion of the tomograms, and ER maps of the cross-sections were displayed using MatLab software (MathWorks, Natick, MA, USA). The schematic diagram of the ERT and corresponding ER diagram grids are shown in Figs. 3 and 4, respectively. D_e was defined as the severity of decay detected by ERT, and it was calculated as (5)${\displaystyle D_e=\frac{R_0-R_d}{R_0}\times 100\text{\%}}$ where D_e was the severity of decay determined by ERT, R ₀ was the average ER value of the detection direction in the section of the same healthy tree species (Ω), and R_d was the average ER value of the detection direction in the decayed section (Ω).

The velocity distribution of stress waves in a cross-section is shown in Fig. 5. The processing of the stress wave velocity distribution was similar to the ER diagram. D_s was regarded as the severity of decay detected by SWT, and it was calculated as (6)${\displaystyle D_s=\frac{V_{\mathrm{j}}-V_{\mathrm{f}}}{V_{\mathrm{j}}}\times 100\text{\%}}$ where D_s was the severity of decay detected by SWT, V_j was the mean velocity of the stress wave in the cross-section of the same healthy tree species (m/s), and V_f was the average velocity of the stress wave in the direction of the decay cross-section (m/s).

Effectiveness of ERT in detecting decay

Used SPSS (SPSS version 19.0; IBM Corp., Armonk, NY, USA) software to perform the regression analysis of D_e and D_t, and the analysis results are as follows.

The correlation coefficient (R²) between D_e and D_t is 0.516 (P < 0.01), and the linear regression equation is (7)${\displaystyle D_e=0.6659D_t+11.852,}$ when D_t was divided into two parts, D_t < 30% and D_t ≥ 30%, some diffidence was revealed. When D_t < 30%, the correlation coefficient (R) between D_e and D_t is 0.677 (P <0.01), and the linear regression equation is (8)${\displaystyle D_e=1.3033D_t+4.2855,}$ when D_t ≥ 30%, the correlation coefficient (R) between D_e and D_t is 0.300 (P <0.01), and the linear regression equation is (9)${\displaystyle D_e=0.7174D_t+8.2687.}$

These results are plotted in Fig. 6.

Effectiveness of SWT in detecting decay

SPSS software was also used to conduct the regression analysis of D_s and D_t, and the analysis results are the following.

The correlation coefficient (R) between D_s and D_t is 0.638 (P < 0.01), and the linear regression equation is (10)${\displaystyle D_s=0.9993D_t+7.5369,}$ when D_t < 30%, the correlation coefficient (R) between D_s and D_t is 0.398 (P< 0.01), and the linear regression equation is (11)${\displaystyle D_s=1.2501D_t+5.9497,}$ when D_t ≥ 30%, the correlation coefficient (R) between D_s and D_t is 0.645 (P <0.01), and the linear regression equation is (12)${\displaystyle D_s=1.3441D_t-8.0242,}$ These results are plotted in Fig. 7

Analysis of the relationship between D_e and D_t

Once the wood is infected by wood-destroying rotten, its cell walls are decomposed and cause the wood to rot and disintegrate. When the wood is rotted and discolored, the hypha growth requires a lot of water, which will increase the moisture content of the decayed area, and then ions will be released from the wood cells. Studies (Houston, 1971) have shown that with the discoloration and decay of standing trees, the content of metal ions such as potassium, calcium, manganese, and magnesium in the rotten wood increase. As the concentration of cations increased, the electrical resistance of decayed and discolored wood was significantly reduced compared to healthy wood (Ostrofsky, Jellison & Shortle, 1997; Nilsson, Karltun & Rothpfeffer, 2002; Jonàs, Carmen & Jan, 2011). The severity of decay detected by ERT (D_e) mainly reflects the increase of the moisture content and metal ions in the decayed trees (Bieker & Rust, 2010). The severity of decay determined by mass loss rate of wood (D_t) mainly reflects the wood weight loss rate, which is closely related to decay distribution range, wood structural damage, and mechanical strength. Both D_e and D_t reflect the different stages of decay in wood. In this study, when D_t < 30%, there was a significant correlation between D_e and D_t, and the correlation coefficient was the highest. Therefore, ERT can make a good diagnosis in the early stages of decay in trees. If ERT is used to detect wood in the early stages of decay, so we can know the condition of the trees early and can deal with the damage caused by decay as soon as possible.

Analysis of the relationship between D_s and D_t

If a large amount of cellulose, hemicellulose, and lignin in wood are corroded by woody rot, decay will occur, and the density will decrease accordingly, and defects will form inside the wood. When the stress wave propagates in the defective wood, it propagates along the edge around the defect. The propagation path changes from a straight line to a curved line. The propagation time increases and the speed decreases (Xu, Xu & Wang, 2014). The severity of decay detected by stress wave (D_s) mainly reflects the size of the internal defects of the standing tree (Tannert et al., 2014), while the mass loss rate of wood is also closely related to the range of decay, the stage of damage to the structure of the wood, and the mechanical strength. Therefore, both methods can reflect the decay status of standing trees, so there is a correlation between them. In this study, when D_t ≥ 30%, there is a significant correlation between D_s and D_t, and the correlation coefficient is higher than the correlation between D_e and D_t. In other words, in terms of decay degree, SWT is a better indicator than ERT when D_t ≥ 30%.

Comparison of two NDT methods

Electrical resistance value of standing trees is affected by many factors such as environmental humidity, temperature, moisture content, the decay stage, growth season, and measurement site, etc., which will affect the measurement effect of the resistance value, moreover, it is easy to misjudge for the sensitivity of resistance testing (Just & Jacbbs, 1998; Wang, Yang & Xu, 2001). Stress wave detection results are influenced by factors such as cross-sectional shape, decayed severity and number of sensors (Gilbert & Smiley, 2004).

In different stages of decay, for these two methods, when D_t < 30%, D_e had a relatively high correlation with D_t. When D_t ≥ 30%, D_s had a relatively high correlation with D_t. These results are related to the decay process of timber. During the early stage of decay, wood quality and visual appearance look unchanged; however, the chemical components have changed markedly. Wood decay fungi can be propagated through the extension and spread of mycelium or mycorrhizae fungi. When wood decay fungi enter wood cells and settle between wood cells, they secrete various enzymes to decompose cellulose, hemicellulose, and lignin in the cell walls of the wood into sugars, which are further digested as nutrients (Van der Wal, Ottosson & De Boer, 2015). Electrical resistance of wood is mainly related to moisture and metal-ion content, so ERT was more accurate during the early stage of decay. Once decomposition of fungi began to stabilize, the variation of electrical resistance became to flatten. During early stage of decay, SWT was inaccurate owing to the lack of cavities (Xu, Xu & Wang, 2014). When cavities were present, SWT was more accurate, since, with the slow growth of holes, the decay grew worse (Li, 2014). Some research has shown that a stress wave could travel around the holes when it encountered them, and then the propagation of the stress wave would lengthen and its travel time increase. According to the calculation method of stress waves, the velocity change of a stress wave could reflect the severity of decay (Tannert et al., 2014).