Abstract
As the intensity and frequency of strong storms increase, the potential for damage to urban trees also increases. So far, the risk of ultimate failure for partially uprooted trees and how they may recover their stability is not well understood. This study sets out to explore if and to what extent trees can regain anchoring strength after their root systems have been overloaded. In 2010, ten London Plane (Platanus × acerifolia) trees were subjected to destructive winching tests. Two trees were pulled to the ground while eight were loaded until primary anchorage failure occurred and were left standing with inclined stems. In 2013, two trees had failed and six were re-tested non-destructively. By 2018, another tree had failed, and we tested the remaining five again. Rotational stiffness was derived for all trials and served as a nondestructive proxy for anchoring strength (R2 = 0.91). After eight years, one tree had regained its original strength, while four had reached between 71 and 82% of their initial rotational stiffness. However, three trees failed during the observation period. The results indicate that partially uprooted trees may re-establish stability over time, but some will not and may fail. In our small data set, it was not possible to identify visual criteria that could provide a reliable indication of tree stability recovery, but our data support the assumption that nondestructive pulling tests can be successfully employed to determine good vigorous candidates for retention after partial uprooting.
INTRODUCTION
The failure of trees with root systems compromised by decay, storm damage, or construction-related damage can pose risk to significant targets and human beings in an urban setting (cf. Smiley 2008; Bergeron et al. 2009; Schmidlin 2009; Smiley et al. 2014; Dahle et al. 2017) and may also pose a risk to those involved in climbing or dismantling trees (Detter et al. 2008). Assessing this structural characteristic of a tree is very difficult. In many cases, when a tree is observed to have significant root issues, the recommendation is to remove the tree. This mitigates risk but also removes the stream of valuable social, environmental, and economic benefits that a tree provides (Price 2007; Roloff 2016; Kim and Jin 2018). It also prevents arborists and researchers from studying whether such compromised trees can recover and regain stability over time.
Root systems are complex subterranean structures that direct a major portion of the wind load collected by the crown into the ground. Below-ground damage to structural roots can often occur due to root decay or root severance and may also be caused by overloading during storm events, by snow loads, or even by heavy impacts (e.g., during road accidents or avalanches).
As the effects of climate change are felt, many predictions indicate that the world will experience more variable and more extreme weather. For example, Cheng et al. (2013) predicted that Canada could receive significantly more wind gusts later in this century and that the magnitude of those gusts would increase. The effect will be stronger for wind gusts over 70 and 90 km/h, and we can expect that more trees will be damaged or destroyed in high wind events. However, not all trees affected by winds experience ultimate failure. It is quite common that after such wind events, some trees are left standing with a lean. While historically many of these trees are removed, there may be alternative management options. Preserving some of these trees may become more important as we strive to increase canopy cover in urban areas for the sake of the benefits provided to those who live among or in close proximity to trees.
Static load tests, as introduced by Sinn and Wessolly (1989), can be effectively utilized to inform tree risk assessments on trees with compromised rooting stability (Smiley et al. 2011; Sani et al. 2012). A tree’s rooting characteristics can be assessed by applying a moderate nondestructive load with a winch, measuring the tree’s reactions with a high-precision inclinometer, and extrapolating those data to determine the minimum strength of the root system (Wessolly 1996; Detter and Rust 2013; Buza and Divós 2016). Estimations of resistance to uprooting are based on comparing this load capacity of the root system with modelled wind loading scenarios for a tree at its actual location, as informed by statistical wind data and local wind conditions (Brudi and van Wassenaer 2002; van Wassenaer and Richardson 2009; Wessolly and Erb 2016; Esche et al. 2018).
The anchorage of trees has been studied in many scientific experiments (cf. for an overview Dahle et al. 2017) and was modelled by several authors (e.g., Dupuy et al. 2005; Rahardjo et al. 2014). Tree uprooting is often described as a progressive failure process that occurs in different stages (O’Sullivan and Ritchie 1993), where a number of components play different roles (Coutts 1983; Blackwell et al. 1990; Nielsen 1991). When the change in stem base inclination does not exceed 0.5° during pulling tests, the process is reversible and nondestructive (Coutts 1983; James et al. 2013). As the stem base inclination increases, the maximum resistance of the root system will be overcome at angles between roughly 2° and 7°; after that point the load applied during the pulling test will decrease as the root-soil matrix progressively fails (e.g., Coutts 1983; Wessolly 1996; England et al. 2000; Jonsson et al. 2006; Vanomsen 2006; Lundström et al. 2007).
Such excessive root plate tilt is likely to cause damage by bending and breaking roots on the leeward side close to the stem and by lifting the windward side of the rootplate, causing horizontal and vertical cracks in the soil as well as bending and ultimately the breaking of roots in tension (Crook and Ennos 1996). When a severe storm partially uproots a vigorous tree, some roots may still be able to retain their water transport function (Ueda and Shibata 2004). Since living wood is weaker in compression, bending failure is initiated by fibre buckling on the compression side (Niklas and Spatz 2014). This fibre buckling may eventually interrupt water transport. However, the fibres on the tension side of mechanically compromised roots and roots less stressed during such catastrophic events may fully retain their water conductivity.
If such a tree is left leaning after the primary anchorage failure, it will usually adapt the orientation of its terminal shoots (Du and Yamamotu 2007) through the formation of tension or compression wood (Archer 1987; Archer 1989). Significant changes in the curvature of the shoot by extension or contraction of the wood tissues has only been observed on stems up to 10 cm in diameter (Berthier and Stokes 2006; Yamashita et al. 2007). It is unlikely that significant changes in shoot curvature will occur on stems much larger than that due to the rapid rise in flexural stiffness with increasing diameter (Fobo and Blum 1985; Coutand et al. 2007).
Furthermore, trees typically respond to a lean by initiating strong increment growth on the side of the stem base in compression from the gravitational loads, which is usually referred to as supporting wood (Götz 2000; Mattheck et al. 2003; Detter and Rust 2018). During wind-induced uprooting, the greatest strains will also occur in the area under compression on the leeward stem base (Stokes 1999). Trees are able to increase increment growth in areas with greater strains (Müller et al. 2006; Larjavaara and Muller-Landau 2010). Finite elements modelling has shown that the addition of wood volume on the compression side of the stem base can be most effective at increasing stability (Yang et al. 2017). The formation of supporting wood at the stem base may be, among others, one mechanism of stability recovery.
Adaptive increment growth is stimulated by a change in the loads that trees experience (Bonnesoeur et al. 2016). For example, healthy forest trees have been found to regain their former stability after a thinning cut within five to eight years (Mitchell 2000). Similarly, after the transplant of both small and large trees, the original root system size could be restored within five to thirteen years (Watson 1985) or sooner (Watson 2005). The effect of root severance on tree stability depends on the distance of the damage from the stem (Smiley et al. 2014), but young trees can recover their anchoring strength as soon as four years after the root severance occurs (Fini et al. 2012).
Our assumption is that trees can recover their anchoring strength within eight to ten years after primary anchorage failure. Experimental data and quantification of stability recovery following overloading of the root system are lacking in the literature. The study presented here provides such data. The degree of root stability recovery after partial uprooting was quantified over a period of eight years.
MATERIAL AND METHODS
All of the research trials described in this paper were undertaken at the Davey Tree research site in Shalersville Township, Ohio, U.S.A., in a plot with London Plane (Platanus × acerifolia) trees. The trees were planted between 1968 and 1970 on Ravenna silt loam. The trials were undertaken in three separate field seasons in 2010, 2013, and 2018. Table 1 summarizes the trees used in the three test series and Table 2 lists their average diameter and height.
Summary of trees used in treatments and as controls in 2010, 2013, and 2018.
Mean height and diameter of trees used in treatments and as controls in 2010, 2013, and 2018 (sd for standard deviation).
The initial research trial was undertaken in 2010. Ten trees with similar diameters at 1 m height, crown shape, and wind exposure were selected for the trial and were pulled until primary anchorage failure occurred. For this project, primary anchorage failure was described as the point during load application (i.e., winching) where the inclination would continue increasing without any further increase in the applied force. This trial could be described as a destructive test since the winching force was applied until the resistance of the root system was overcome (treatment groups). Once these original destructive pulling tests were completed, two of the trial trees were subsequently pulled to ultimate failure, i.e., until the trees uprooted completely and fell to the ground (treatment group 1). Eight trees were left standing on the site (treatment group 2).
At the same time, in a separate experiment on the same site with the same tree species, a second set of ten trees was pulled non-destructively to 0.25° of inclination at the root plate before they were pulled to ultimate failure. The force and inclination data gathered from the nondestructive portion of those trials (at 0.25° of inclination) was used as a control reference for the rotational stiffness of trees on the site that did not experience primary anchorage failure in 2010 (control group 1).
The second field season was in 2013 at the same site. The six remaining trees were subjected to a nondestructive pulling test with the same configuration as the previous trials. The trees were pulled to 0.25° of inclination in the same direction and to the same anchor points. In 2013, another trial was also undertaken on the same London Plane plot. Six new trees were pulled non-destructively to 0.25° of inclination at the root plate before they were pulled to ultimate failure. The force and inclination data gathered from the nondestructive portion of those trials (at 0.25° of inclination) was used as a control reference for the 2013 pull tests of trees from treatment group 2.
In 2018 the site was revisited for a third time. The remaining five trees were retested non-destructively and pulled in the same direction to the same anchor points. Five other London Plane trees that were in the same plot and had not been winch-loaded in any of the previous trials were selected as controls for the 2018 trial. These trees were all pulled non-destructively to a maximum of 0.25° of inclination using the same protocols as the previous trials (control group 3).
While the winching tests were underway, the applied load was measured continuously with a forcemeter (load cell) in the pulling line, and the resulting root plate rotation was measured with two bi-axial inclinometers (one at the side of the stem base, one at the back). The instruments used are part of the TreeQinetic system (Argus Electronic GmbH, Germany). Inclinometers had a resolution of 0.001° (accuracy 0.002°) and the forcemeter had a resolution of 0.1 kN (accuracy 0.3 kN). The rope angle from the horizontal was measured by using a digital level (Digipass, United Kingdom) with an accuracy of 0.2°.
The test was configured according to the Static Integrated Method or Pulling Test Method (Sinn and Wessolly 1989; Brudi and van Wassenaer 2002). The applied force was converted into its lateral component by the cosine of the rope angle. The bending moment was determined as the product of the lateral force component (in kN) and the lever arm length as the vertical distance from the stem base to the anchor point of the rope (in m).
Rotational stiffness at the stem base was calculated for all trees in our data set as the bending moment at 0.25° of basal inclination and served as a nondestructive proxy for anchorage strength. Anchorage strength was defined as the maximum bending moment that occurred during the winching tests. It was only measured for trees that we pulled to primary anchorage failure. In order to account for differences in tree size, rotational stiffness was scaled by tree size (height × diameter2) when different groups were compared with each other. Data were analysed with a random slope and intercept linear mixed effects model (Pinheiro and Bates 2000) adjusting for variance between years using the statistical analysis software R (R Core Team 2018).
RESULTS AND DISCUSSION
In our data set, measured rotational stiffness proved to be a good indicator for a tree’s anchorage strength (Figure 1). This is consistent with the findings in earlier experiments (e.g., Wessolly and Erb 1998; Brudi and van Wassenaer 2002; Jonsson 2007; Smiley 2008; Detter and Rust 2013). The correlation is very strong in our specific data set (R2 = 0.91), presumably because all of the pulling tests were undertaken on the same plot of land, and because we used trees of the same species as well as of similar age and size (cf. Table 2). Limitations to this approach could result from drastic changes in the soil water content (Kamimura et al. 2011) which we avoided in the present study. Minor changes in the soil moisture content will not considerably affect the rotational stiffness (Rust et al. 2013). Therefore, it was possible to determine the current anchorage strength of trees in our data set using data on their rotational stiffness that was gathered during nondestructive tests. This allowed for the assessment of anchorage strength recovery among the remaining trees from treatment group 2 without pulling the subject trees to failure a second time.
Rotational stiffness indicates anchorage strength (adjusted R2: 0.91). This figure contains all the trees from this study that were pulled beyond primary anchorage failure.
The data for the study trees showed similar load vs. inclination curves to those observed in earlier uprooting experiments (e.g., Coutts 1983; Wessolly 1996; Lundström 2007; Detter and Rust 2013; Buza and Divós 2016). After primary anchorage failure during the first treatment in 2010, the trees remained leaning by more than 2°, except for #277, the only tree that was not loaded beyond 1.2° in the initial winching test (Figure 2).
Load vs. inclination diagrams for both treatment groups in 2010.
The two trees in treatment group 1 (#284 and 285) showed a reduction in rotational stiffness of 44 and 56% respectively between the first test undertaken to primary anchorage failure and the second winching test directly after the first. Once the load at which the first test was terminated was exceeded in the second test, the load vs. inclination curve continued almost as if the tree had been pulled to ultimate failure all at once (Figure 3). We suspect that the root system was damaged by the first winching test, but the damage was restricted to a certain level since the uprooting process was interrupted. As winching was continued beyond this point in the second test, the root system may have resumed the progressive failure process until the tree was on the ground. This observation was the subject of further studies undertaken by the authors during the Tree Biomechanics Research Week in 2013, but those studies are beyond the scope of this paper.
Load vs. inclination diagrams of treatment group 1. The first winching test to primary anchorage failure is shown on the left, and the second test to complete uprooting is shown on the right. The horizontal lines mark the load at which winching terminated in the first test and where the progressive failure was resumed in the second. Data recording was automatically terminated at roughly 30°.
For treatment group 2, two trees had failed in the three years since the original trial, and six trees remained standing in 2013. Out of those six trees, one more had failed in the subsequent five years, leaving five of the original trial trees standing in 2018. The load vs. inclination data for different years is shown in Figure 4. A linear mixed effects model statistical evaluation gives P = 0.0016 for the difference between 2010 and 2013, and P = 0.0815 for the difference between 2010 and 2018. The pronounced difference between rotational stiffness in 2013 vs. the first treatment in 2010 indicates that the initial winching had significantly damaged the root system. However, after eight years, the load response of all remaining trees was not significantly different from their predamaged responses in 2010.
Load vs. inclination data from treatment group 2. From the winching test in 2010 (left), only the nondestructive part up to 0.25° is displayed. For 2013 (middle) and 2018 (right), the first loading cycle up to 0.25° inclination is displayed.
The rotational stiffness was derived from the original data shown in Figure 4. Changes in rotational stiffness between years are shown in Figure 5. After three years, the rotational stiffness of four trees ranged between 44 and 66% of the rotational stiffness found when the trees were undamaged in 2010. After eight years, the rotational stiffness of those four trees was still less than the rotational stiffness found in the 2010 tests (71 to 82%). None of those trees reached their original rotational stiffness within the observation period, but none of them failed during that eight year period either.
Rotational stiffness in years 2010 (left column), 2013 (middle column), and 2018 (right column) among the trees in treatment group 2, standardized to the initial winching test in 2010.
A fifth tree (#278) failed somewhere between 2013 and 2018. This tree had lost two thirds of its original rotational stiffness by 2013, indicating that it never recovered after the initial damage in 2010. At the same time, the rotational stiffness of the sixth tree (#277) was only reduced by roughly 10% in 2013, and by 2018 its rotational stiffness was higher than the value recorded prior to the 2010 treatment. This result can likely be explained by the fact that the anchorage of tree #277 had not been compromised as seriously as other trees during the initial winching treatment of 2010.
In the control groups, different trees were tested each year. In order to enable meaningful observations, we scaled the measured rotational stiffness by tree size (Figure 6). It is not surprising that there are obvious differences between the rotational stiffness of the control trees in different years, because these were different trees. Nevertheless, the range of values was similar for treatment groups and control groups in 2010 and 2018, respectively. Only in 2013, the rotational stiffness of treated trees (treatment group 2) fell drastically below that of the control trees (control group 2).
Boxplot of rotational stiffness scaled by tree size. In the treatment groups (left), the same trees were tested, but the control groups (right) consisted of different trees each year. In both cases, the sample size fell from N = 10 in 2010, to N = 6 in 2013, and N = 5 in 2018.
The rotational stiffness of the trees in treatment group 2 clearly shows a decrease from 2010 to 2013 and an increase from 2013 to 2018. The former may be the consequence of the damages generated from the destructive winching tests, while the latter can demonstrate the subsequent recovery of the strength of the trees’ anchoring systems and confirms our hypothesis. Since Figure 6 shows no such trend for the control groups, we are confident that the observed effect within treatment group 2 did not occur due to either climatic or seasonal effects, nor is it likely the consequence of generic circumstances that would have affected all trees on the test site.
It is likely that only healthy trees will be capable of making the adaptations required to regain their anchorage strength after significant damages have occurred. The vitality (assessed visually) of the trees tested in this experiment was generally good and did not vary over time. The number of trees in this study was very limited, and only one species was studied. The findings may differ for trees with lower vitality, trees growing under adverse conditions, larger or smaller trees, or trees of other species. Therefore, additional studies are required to enlarge the empirical basis for quantifying anchorage strength recovery after primary anchorage failure.
When arborists visually inspect trees with increased leans, they should be able to recognize symptoms of root failure after significant storms and also draw conclusions from signs of growth adaptation in response to earlier events (Dunster et al. 2013; Smiley et al. 2017). The results of this study indicate that some trees are capable of recovering their stability over time after primary anchorage failure has occurred. Therefore, some insights into the current stability of leaning trees can be made by assessing current tree vitality, a tree’s self-correcting response, and the formation of supporting wood (Detter and Rust 2018). Since the likelihood of ultimate failure is generally higher for partially uprooted trees in urban situations, visual assessments alone may not be sufficient to identify which trees are good candidates for retention.
Three out of eight trees in this study that were left standing after the initial winching tests subsequently failed within eight years, while five others recovered most of their original stability during this period of time. Tree pulling tests can be used to effectively determine a conservative estimate of a tree’s ability to withstand strong wind events. The pulling test results for tree #278 in 2013 showed an exceptionally high loss of rotational stiffness. This loss of stability could have been detected with a pulling test during a level three tree risk assessment (Smiley et al. 2011) and mitigation could have been prescribed if a target would be affected by a failure. Since this tree failed between 2013 and 2018, the nondestructive pulling test had correctly identified its weakness.
Finally, the study shows that some trees can survive partial uprooting, presumably by correcting their growth direction, formation of supporting wood, regrowing roots, and thus eventually restabilizing after a period of time. The pulling test method can help to measure and quantify this effect non-destructively and could be used in conjunction with preventive guying to help identify and preserve some partially uprooted trees rather than removing them. The remaining trees from treatment group 2 may be retested at a later date and eventually harvested to study the strategies of morphological adaptation within their stems and the actual mechanisms involved in the recovery of anchorage strength.
ACKNOWLEDGEMENTS
The authors would like to thank the ISA, the TREE Fund, Davey Tree, and all the volunteers for their generous contributions to make Tree Biomechanics Week a success and for their logistical support of this research. The research could never have succeeded without the amazing contributions of following field technicians: Alex Satel, Matt Follett, Will Koomjian, Mike Neuheimer, Josh Galiley, Taylor Hamel, Chris Cowell, Benedikt Morbach, and Ryan Lewis. We also would like to thank Alex Satel for reviewing an earlier version of this paper and providing helpful suggestions. Furthermore, Gary Watson and Jake Miesbauer provided comments and helped to improve the paper. Part of this work was funded by the German Ministry of Education and Research (research grant 17021X11).
- © 2019, International Society of Arboriculture. All rights reserved.
LITERATURE CITED
Jump to section
Related Articles
Cited By...
- No citing articles found.