Patterns of Structural Failures in Urban Trees: Coast Live Oak (Quercus Agrifoua)

Methods

The California Tree Failure Report Program survey form and the data categories have been described previously (Costello and Berry 1991). Standardized report forms are completed by tree workers, arborists, or consultants encountering a structural tree failure and are submitted to the CTFRP. The data were compiled with dBaselV database software. Coast live oak failure reports were analyzed using the dBStats (Ashton-Tate) statistical package. Custom applications were written for dBStats to perform repeated comparisons between species based on various data fields, logical tests for the creation of screened fields, and field and value labeling for printout readability. Data from selected tree failure report form fields were grouped into categories to facilitate specific comparisons. For example, a new field was created by the software application to divide data from the “location of failure” field into the general categories “above ground” (trunk failures above ground level) and “at or below ground level” (trunk failures at ground level and root failures) to compare soil-level problems to above-ground-level problems. Missing information was screened from individual fields and removed from statistical analysis.

For general data descriptions, frequency tables (percentage of population) were used directly or in combination with sorting features to establish additional categories (such as height classes or “other oaks”). Crosstabulation of fields (i.e., tree group vs. failure location) was analyzed using the χ² test on 2 × 2 or 2 × 3 contingency or crosstabulation tables (Geng and Hills 1989), where P indicates the probability that compared groups share the same distribution of values. Where data for tree age are used, it should be noted that in most cases age was estimated to the best of the ability of the reporter and that accurate age estimation is often difficult due to species and site variability.

Comparisons with 5 other major groups in the CTFRP database were useful in defining the characteristic failure patterns of coast live oak. For the purpose of comparison, a group may be defined as a single species, such as valley oak (Q. lobata); as a grouping by genus (Eucalyptus); or as a larger grouping of many genera, for example, all hardwoods or all conifers. Such defined groups are most useful if they are reported frequently in the database, as is the case for valley oak (5% of the total), and eucalyptus (12.6% of the total). We have separated eucalyptus from the hardwood category because it exhibits failure characteristics more similar to conifers than to other hardwoods.

Results

Coast live oaks and other oaks in general are typically trees indigenous to the site (80% of coast live oak reported). In contrast, approximately 90% of other hardwood species, eucalyptus, and conifers in the database are planted trees. Coast live oak failures were reported most frequently (57%) from residential sites. Eucalyptus and conifer failures were commonly reported from parks (39% and 56%, respectively). Failures of “other hardwoods” were reported most often from residential (36%) and street tree (28%) sites.

Coast live oaks, valley oaks, and “other oaks” (the remainder of species excluding Q. agrifolia and Q. lobata grouped together) failed at a significantly higher reported age (110, 110, and 156 years mean, respectively) than other hardwoods, eucalyptus, or conifers (33, 40, and 60, respectively). The mean reported age of valley oak (156 years) was significantly higher than all other tree groups. When comparing mean dbh (diameter at breast height), coast live oaks (88 cm [35 in.]), other oaks (80 cm [32 in.]), conifers (88 cm), and valley oaks (108 cm [45 in.]) were significantly larger than other hardwoods and eucalyptus (50 and 65 cm [20 and 26 in.], respectively). Valley oak was significantly larger than all other tree groups.

Location of Failure

The most general structural characteristic of any tree failure is where it occurs on the tree, whether on the trunk, branches, or at the roots. In Table 1, the relative distribution of failure locations is compared among the 6 tree groups. These distributions were tested for significant differences by pairs of groups in a 2 (tree group) × 3 (location of failure) cell matrix using the χ² test. Coast live oak failed most frequently at the root system (45% of reports). However, the χ² probability was not significant between coast live oak and other hardwoods (0.13), valley oak (0.16), or other oaks (0.36), even though these groups had a relatively more uniform distribution of location of failure. Eucalyptus and conifer species shared a common distributional pattern for location of failure (χ² P = 0.44) with a high percentage of branch failures. Both eucalyptus and conifer location of failure distributions were significantly different from the hardwood group and from coast live oak (χ² P< 0.001).

Another way of looking at the location of failure is to group failures into those occurring at the ground level or below (including ground level trunk failures and root failures) and those occurring above ground level (trunk failures occurring above ground level and branch failures). Using these groupings (Table 1), the failure pattern of coast live oak (46% above ground level, 54% at or below) is significantly different than that of hardwoods (χ²P = 0.07), and valley oak (χ² P = 0.02).

Decay

Coast live oak, like other oaks in the database, had a higher incidence of decay associated with failure compared with the remaining hardwoods in the database, with conifers, or with eucalyptus. Table 2 shows the frequency of decay associated with failure for these tree groups, separated by failure location. “Decay” combines together all decay types (heart, sap, etc.) reported as associated with failure under “decay type” on the tree failure report. Chi² probabilities are assembled in Table 3 for 2 × 2 matrix comparisons between pairs of tree groups from Table 2 and the decay index (presence or absence). The “all oak” group pools all oak species.

Decay was associated with failure in oak more frequently than in other hardwoods (e.g., 83% of all failures for coast live oak). Individually, there appear to be differences among the oaks. Coast live oak had significantly higher association of decay with failure than other hardwoods except at the root. Coast live oak had a lower association of decay with failure at the root than other oaks. Coast live oak had a significantly higher decay associated with trunk failure than Q. lobata.

Structural Defects

Tree structural defects are potentially important factors contributing to a failure. The California Tree Failure Report Program includes a category for recording structural defects, such as multiple trunks/ codominant stems, dense crown, leaning trunk, cracks/splits, kinked or girdling roots, and others. Table 4 includes the defects that had the 3 highest values for percentages of failure within each of the 6 tree groups. The most frequently occurring structural defect/failure location combination for coast live oak was “failed portion dead” at the root. The most frequently occurring structural defect/location combination for all tree groups other than coast live oak was a branch failure associated with heavy lateral limbs. For conifers, branch failure was an exceptionally high percentage (36.5% of all failures); for eucalyptus and “other oaks” it was a high percentage (about 18%), and for hardwoods and valley oaks it was about 10%, slightly higher than the next most frequent failure combinations.

The structural defect “failed portion dead” was commonly associated with both root and trunk failure of coast live oak, together accounting for 19% of reported failures. It did not show up as a major factor for eucalyptus or conifer failures. Hardwoods, other oaks, and valley oaks also had a lesser percentage of failures associated with dead portions of trunk or roots (6.5%, 11.4%, and 8.2%, respectively).

Environmental Conditions at Time of Failure

Environmental conditions occurring at the time of a tree structural failure may be a contributing factor to that failure or may indicate threshold conditions for a particular species or tree group. The CTFRP report form has fields for reporting wind speed and direction, precipitation, temperature, soil condition, and site topography at the time of failure.

Discussion

Overall, the most common combined pattern of structural failure for coast live oaks reported in the CTFRP was a root or low trunk failure of a naturally occurring tree growing in a residential site (Figure 1). Coast live oaks reported in the CTFRP failed most often at the root (45% of Q. agrifolia). When trunk failures at ground level were combined with root failures, coast live oak had a significantly higher frequency for ground-level and below-ground failures than other tree groups (Table 1). This places the major failure problem area for coast live oak at the soil level and below, suggesting that human site and soil modification can have a negative impact on the health of these remnant trees of native origin. Site grade changes, both cutting and filling, were frequently associated with the failures of native oaks. Coast live oak in particular has the highest frequency of grade change associated with failure (18% fill and 12% cut). Reduced gas exchange under compacted soil or fill conditions has been shown to correlate with decline symptoms in oaks (MacDonald et al. 1993). Low oxygen conditions reduce root growth and increase disease severity in oaks (Jacobs et al. 1997a, 1997b). Cut roots as a result of grade modification operations may also expose entry avenues for disease organisms.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Failure of Q. agrifolia in San Luis Obispo County, California. The tree had a dbh of 100 cm (40.5 in.). It failed at the roots during heavy spring rains, with soggy soil and gusty winds. The failed tree had been growing in a residential setting that was a native stand prior to human occupancy (visible in the background). Note the soil disturbances: Herbaceous ornamentals were planted around the base of the tree, requiring irrigation on the trunk and root system; and vehicle traffic, parked cars, etc. extended almost to the trunk. Root decay was present. This photo was taken by CTFRP cooperator Ted Elder 1 to 2 months after the failure.

Saturated soil conditions associated with root failure were also more frequent for coast live oak failure than for other tree groups. Soil saturation also reduces soil gas exchange and may thus lead to root decline, although it is not known how long the soil was saturated for each individual failure. Saturated soil conditions impart less soil structural stability as well, placing greater stress on weakened or dead root system components.

Decay and dead wood were significantly more frequently associated with coast live oak failures than with failures in other tree groups. Eighty-three percent of failures were associated with decay in coast live oaks, similar to the decay pattern observed in other oaks (86%). Decay is thus a major contributing factor to tree failure in coast live oaks. Decay data are hard to interpret. While oaks in general have a higher association of decay with failure than other tree groups, they also fail at a higher reported age than other tree groups and at a larger dbh by a factor of at least 2. While tree age may be a questionable measure due to the imprecision of visual estimation, the older nature of the oak population seems to be reflected in the dbh data. Published dbh tables for Q. agrifolia were unavailable; however, dbh tables for Q. kelloggii indicate that a 88-cm (35-in.) dbh (the average dbh for Q. agrifolia failures) corresponds to an age of 200 years (Burns and Honkala 1990). The oak population’s apparent older age before failing may suggest that oaks are actually more resistant to decay rather than more susceptible, or that oak wood is of higher strength and can withstand a greater loss of sound wood before failing. There are few data regarding the percentage of decay in the wild for coast live oak or regarding the age of onset of decay in oaks that have failed.

A large proportion of coast live oak failures (49% of failures overall) occurred in conditions where wind speed was less than 5 mph (8 kph). At a lower wind speed, trees experience a lower bending stress. Our finding that these trees frequently fail at lower wind speeds, together with decay and structural defect data described above, suggests that a substantial decline in structural strength had already occurred in coast live oaks before the trees failed.

Management Strategies

Our findings confirm that, to maintain the health of coast live oak and California oaks in general, careful attention to controlling excess soil moisture and avoiding or correcting grade disturbance and compaction is essential, as is monitoring decay throughout the tree, especially in below-ground portions (Britton 1990).