We evaluated the structural response to simulated wind stress applied to two coastal tree species native to the southeastern USA that have contrasting rates of windthrow during hurricanes: Baldcypress (Taxodium distichum) and laurel oak (Quercus laurifolia). In a static winching study, we evaluated the stability of 30 individuals each of laurel oak and baldcypress by quantifying the critical turning moment required to cause structural failure. Biomass of trunk segments and branches was measured directly in the field for a subset of trees pulled and used to develop regressions to determine biomass of other trees. Baldcypress were more likely to snap than uproot, while the opposite was true for laurel oak. Baldcypress critical turning moments were higher than laurel oak based on biomass, but were similar when compared by diameter and other size proxies, and ranged from 10.9–43.9 kN m. Models that incorporated crown size and shape indicated smaller baldcypress could withstand higher wind speeds than laurel oak before failing. Unlike laurel oak, many smaller baldcypress were so flexible that they bent to the ground instead of failing. We conclude that baldcypress will remain wind resistant and thus ecosystem services and functions provided by this species should be maintained. Conversely, laurel oak, while stable at large diameters, is more likely to experience catastrophic failure and be less stable at smaller diameters than baldcypress, which suggests that their presence in built environments should be carefully evaluated.

We selected 30 baldcypress and 30 laurel oak throughout seven field sites located in Hobcaw Barony and the Clemson-Pate Forest, just east of Georgetown, South Carolina (Lat. 33.361308°, Long. -79.224524°). Stem diameters ranged from 10-47.9 cm and were measured at breast height (1.37 m) for laurel oak, and above buttress for baldcypress. We selected individuals without obvious defects in their trunks that could affect structural integrity.

We used the same winching setup and calculations as in other studies (Fredericksen et al. 1993; Nicoll et al. 2006; Peterson and Claassen 2013; Cannon et al. 2015) except that we used accelerometers per Garms & Dean (2019) instead of inclinometers or dynamometers (see publication). Trees were scaled using a self-ascending deer stand. Strap attachments on winched trees were placed at ~1/3 the height of the tree, depending on branch locations that prevented further ascension by the climber. We used a Badland Apex 12000-lb (5443 kg) winch with synthetic HMPE rope (Dyneema ®) rather than wire rope or cable used in older winching studies because of its superior strength and lighter weight. We used appropriate personal protective equipment (PPE), Bluetooth sensors, and a remote-control to operate the winch from a safe distance. Field gear was inspected each day and replaced as needed to ensure personal safety.

A Crosby SP Radiolink Plus load cell with Bluetooth connection to a cell phone (Crosby|Straightpoint, 2023) was attached between the mechanical winch (or pulley if needed) and the strap attached to the study tree to measure force. One accelerometer was placed just below the strap attachment and one at the base of the tree. The accelerometer heights were recorded after placement. The timestamp at peak force measured on the load cell was matched to the accelerometer’s timestamp for tilt to a tree’s moment of failure (“critical tilt”). The critical tilt in radians () of both accelerometers was averaged and used to determine the height of strap attachment above the ground and the horizontal displacement at strap height during the moment of trunk failure.

The mode of failure of a study tree was recorded using the following categories: snap, uproot, splinter (trunk bent and splintered but did not completely snap and did not uproot), bent (no snap, no uproot, trunk did not splinter, bark remaining mostly intact, tree partially rebounded when released), and bend-snap when trees did not snap until being pulled > 60° from vertical.

Total tree biomass and vertical center of mass were calculated to determine the vertical force contributing to failure by a tree’s self-weight. Tree height was recorded with a measuring tape on the ground after winching. Three trees of each species (“biomass trees”) were cut above a 20 cm stump, which included most of the buttress of baldcypress, and weighed. Branches with leaves were weighed the day of felling. Trunks were cut into ~1 m segments and weighed in the field with a hanging balance within a few days to minimize water loss. The precise length of each segment and its diameter at the top and bottom were measured to determine volume of the segments. We used these data to determine tree volume with the formula for a truncated cone and then developed regressions to determine the biomass of other laurel oaks (Mass (kg) = 0.000836 * volume (cm³); R² = 0.99) and baldcypress (Mass (kg) = 0.000895 * volume (cm³); R² = 0.98) used in our study.

To create an accurate regression to determine branch biomass for use on other trees, we measured fresh weight in the field of the branches ≥ 2 cm diameter from the biomass trees and additional branches from other felled trees to span the range of branch sizes present. Because baldcypress and laurel oak have different branch structure, we used different formulas to best capture branch biomass. We calculated laurel oak branch volume with the formula for a cone using branch basal diameter and branch length and from this developed a regression equation (Mass = 0.001814 * Volume; R² = 0.9034) used to estimate branch weights of other laurel oak branches (Figure S2). We used a similar process for baldcypress (Figure S2) but because it has smaller, less structurally complex branches, biomass was accurately estimated with branch basal diameter (d) alone (Mass = 0.1917^2.0415; R² = 0.7381).

Total tree mass was determined from the sum of each segment’s mass and the mass of each branch on a study tree, or by applying the derived allometric equations. Branch weight was added to trunk weight of the corresponding 1-m segment. For branches that were not oriented horizontally, we partitioned their weights among the corresponding trunk segments going up the tree. The biomass data were used to determine tree center of mass (x_m) to determine the self-weight and gravitational components of M_crit.

Calculation of Mcrit is well established in the literature (Fredericksen et al. 1993; Nicoll et al. 2006; Peterson and Claassen 2013; Cannon et al. 2015). The rotational force at the moment of tree failure is quantified as the critical turning moment (M_crit, measured in Newton-meters; N-m) and can be used to estimate wind forces needed to topple trees. The critical turning moment is determined using a static winching procedure and is calculated by summing the maximum turning moment (M_applied) measured by the load cell, plus the gravitational force contributed by the self-weight of the tree (M_self).

M_crit = M_applied + M_self

Step by step calculations can be found in Supplemental materials from the accompanying article.