Highway repair and its impact upon tree health

In urban areas, street trees will normally exist within a length of grass verge (of varying width) running parallel to the highway (path and / or road), or within planting pits inside the highway. As these trees mature, both the roots and the root collar can cause damage to the highway – particularly if rooting space is limited. When such damage occurs, there is generally the need for remedial works to take place. However, such remedial works can have an impact upon tree health, survival, and economic value (CAVAT, CTLA, and so on) – particularly as roots may be damaged or severed during the construction works, which has consequences for tree health.

An example of a street tree drastically uplifting the pathway and curb. Source: Seattle.gov.

In light of the above, research was undertaken during the late 1980s to early 1990s in Milwaukee, USA to establish exactly the above impacts to trees that have had highway repair works undertaken within their rooting environments. The reason for Milwaukee being the choice location was because the city had a third of its trees valued with the CTLA system in 1979, so there was relatively recent data to compare results to with regards to economic impacts.


The authors looked at construction schedules for the 1981-1985 period within the areas where trees were assessed with the CTLA valuation system, and pinpointed locations where highway repair or widening had been undertaken (in order of descending ‘disturbance severity’, the four criteria established by the authors were: street widening and curb setback, curb and pathway replacement, curb replacement, and pathway replacement).

One hundred projects were then randomly selected over the 1981-1985 period, with 20 per year (allowing for tree condition to be assessed 4-8 years after highway repair works were undertaken). From each study, a single block where the repair works took place was identified. Then, using the criteria mentioned above, 50 blocks (10 per year) were selected based on the highest ‘disturbance severity’. These 50 blocks were then examined, and the top 25 in terms of species diversity were selected to feature within the study. The authors then chose the nearest block to those 25 blocks that did not have repair works undertaken and contained tree populations, and used them as the controls. The first 25 trees in both the construction and control blocks were then identified (from the 1979 survey) for surveying. This lead to 989 trees being sampled in total – 510 from construction blocks, and 490 from control blocks.

From each tree, the following data was collected: DBH at 1.4m, species, verge width (from pathway edge to curb), and the CTLA tree condition rating (100, 80, 60, 40, 20, and 0). The type of construction activity was also identified. Then, comparisons were drawn to the 1979 CTLA valuation survey, to determine whether the trees, assessed in 1989, had suffered as a result of repair works between 1981-1985.


Of the 989 trees sampled, only 670 had actually survived from 1979-1989. 175 trees had been replaced since 1979, and another 144 were newly planted in different locations. The trees were of 15 different species, though Acer platanoides, Fraxinus pennsylvanica, and Gleditsia triacanthos were the only three to feature enough to have statistical analyses run on them.

With regards to tree condition (with the percentage being the average of the entire CTLA tree condition rating scores, I suspect – it’s not clarified), no significant difference was found between the construction (77.2%) and control (77.7%) blocks in the 1979 survey, though by 1989 there was a significant difference between construction (71.2%) and control (76.7%) blocks.

Table taken from the report.

In terms of tree survival between these years, 81.4% of the trees on control blocks survived, whilst only 77.3% survived on the construction blocks – again, a significant difference.

Table taken from the report.

In relation to verge width, significant differences were again found. In both control and construction blocks, a lower width resulted in trees being poorer in condition, though where construction had occurred the decrease in condition was more distinct.

Table taken from the report.

Significant difference in tree condition between construction and control blocks was also observed between tree species (in this case, only Acer platanoides, Fraxinus pennsylvanica, and Gleditsia triacanthos could be analysed, as other species were not in enough abundance), using two-way ANOVA. However, using one-way ANOVA, there was no significant difference.

No significant difference between construction and control blocks was found with regards to tree diameter.

What does the data suggest?

The authors begin by asserting that highway repair works has a significant impact upon both tree condition and survival. For instance, a 22.7% mortality rate was observed in trees on the construction blocks, compared to 18.6% on control blocks. Similarly, whilst the condition of control trees did not significantly change during the survey period, it declined by 6.1% for trees affected by construction. Results also suggest that the width of the verge has a direct impact upon tree condition on both construction and control blocks, though trees on narrow verges that also were impacted by construction suffered more significantly.

How will this tree be impacted by the highway repair works direction surrounding it? Source: Bike Nopa.

Conclusions drawn from the data also suggest that tree species is not a significant determinant in tree condition and survival rate following construction. The authors state that this was to be expected, because all three species aforementioned are hardy species that are tolerant of urban conditions and disturbance. In terms of tree size (DBH), the authors note that they were surprised no significant difference was found between control and construction blocks, though because many of the trees were young (319 were under 10 years of age, and many more were planted in the 1960s and 1970s), this may have impacted upon the data. If all, or more, of the trees were mature, it may have been a different story entirely.

Turning attention towards economic implications of highway repair works to trees, assuming each tree was worth $1,100 (as was, I suspect, concluded in the 1979 survey), the 200,000 trees of Milwaukee would value in at $220,000,000. As around 3% of the tree population had a decline in condition resulting from highway repair works each year during the study period of 1981-1985, a total of 6,000 trees per annum (with a value of $6,600,000) would suffer, meaning an annual loss of $521,500 can be calculated. Additionally, as tree mortality associated with construction works was 4.1% higher than for control blocks, an additional hit of $270,600 would be taken. Therefore, the effect of highway repair works on the value of Milwaukee’s tree population was $792,100 per annum, between 1981-1985.

Source: Miller, R. & Hauer, R. (1995) Street Reconstruction and Tree Decline. In Watson, G. & Neely, E. (eds.) Trees & Building Sites. USA: International Society of Arboriculture.

To discuss this post, please either post here or head over to Arbtalk and discuss it there.

Highway repair and its impact upon tree health

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s