Impact of Commercial Development on Austin's Urban Forest



Background Information:

This page provides an exposition and analysis of apparent trends regarding impacts to Austin's urban forest from commercial development. The information used in this analysis comes from tree removal and replacement data collected by the City of Austin as part of the development review process. The City of Austin records this data for all significant and protected size trees on a proposed development site, and it is used to help enforce the City's Tree Ordinances
Recorded data is in units of summed diameter inches as measured at breast height (4.5 feet above the ground). These measurements will simply be referred to as "inches" throughout the page. 

For each development site, records are kept on the sums of all surveyed inches, removed inches, and mitigated (replanted) inches.

To illustrate by example,  suppose a development site has two (2) twenty-inch trees and seeks to remove one. Suppose further that the City has required one tree of at least 5 inches to be replanted as mitigation for the removal. The tree data recorded for this site, then, would be 40 inches surveyed, 20 inches removed, and 5 inches mitigated. 

Briefly, it is worthwhile to examine what is meant by commercial development.
Commercial development is understood here to consist of two broad categories: sites plan and subdivision.
For the sake of this analysis,  a site plan is understood to include any proposed development site that is not a single family residential home or duplex. Boat docks, though required to undergo site plan review, are not considered commercial in nature.
Similarly, not all types of subdivisions are taken into consideration. Emphasis will be on subdivision construction (e.g. road and drainage construction), as it is during this phase of the subdivision process that actual impacts to the urban forest can be expected. 

As a final note, the data underlying this analysis is updated daily as new developments sites are approved. Furthermore, every map and figure on this page is interactive. Please feel free to use the interactive features to explore the data presented. The full data set can be downloaded here.

Introduction:

Commercial development has a significant impact on Austin's urban forest. Despite its tree ordinance, Austin still lost approximately 505 acres of tree canopy on commercial development site plans between the years of 2010 and 2014, presumably as a direct result of tree removals for construction or other development related impacts to trees. To put this figure into perspective, that is an area more than 10 times greater than Auditorium Shores or 1.5 times the area of Zilker Park.   
Tree removals from commercial development have the potential to significantly reduce Austin's overall tree canopy over time, which could, in turn, exacerbate environmental problems such as flooding and the urban heat island effect.  It is, thus, important to understand what trends are evident, so as to devise effective strategies for preserving and growing Austin's valuable urban forest.   
To put the matter into perspective, the time lapse image below shows the effect of all development for the last 33 years.  

Analysis of Temporal Trends:

Austin has seen a steady increase in the number of commercial development sites since 2012. With the most significant yearly increase occurring in 2014. That year saw an increase of 27% from the previous year. 
Despite the yearly increase in the number of construction sites, tree removal activity does not follow the same pattern.  As can be seen in the chart below, there are only occasional large spikes in tree removals. For instance, while several months between January of 2014 and August of 2015 show exceptionally heavy tree removal activity, the monthly sum of removed inches generally
tends to stay under 15,000. 
The spikes in total monthly tree removals are not necessarily caused by the increase in development sites. Rather the spikes are attributed to just a handful of large development sites that remove more trees than what is typical. 

Tree Removals (Moving Average)

Moving averages are commonly used with time series data to smooth out short-term fluctuations and highlight longer-term trends or cycles. The chart below gives is a 30-day moving average. It gives a better picture of what is typical in terms of monthly tree removals from commercial development. 
Overall, tree removals do not appear to be increasing. Rather it appears that 2014 and 2015 saw an exceptionally high level of tree removal activity. 

Outliers in the Data

The chart below gives a sense of the overall spread of the data. Take particular note that the scale of the y-axis is logarithmic, instead of linear (i.e. it increases at each interval by an order of magnitude). A strong positive skew to the data is apparent.
Whereas a typical removal site takes out about 152.5 inches (median), there are several extreme cases where thousands or even tens-of-thousands of inches were removed. The most extreme case had nearly 49,000 inches removed. 

The Disproportionate Effect of the Outliers

The top ten removal sites since 2012 account for approximately 25% of the total removal for the same period. The largest removal site, the Avery Station subdivision, accounts
for ~10% * of the total removals. Aerials for this site before and after development are displayed below. Details of the project can be found here.

Due to the presence of large removal sites, such as what is pictured above, and a majority of small removal sites, the relationship between the sum of removed diameter inches and the overall number of development sites approximates the Pareto distribution*, wherein the majority of trees being removed are coming from a minority of the cases overall. This skewness helps explain the lack of a strong correlation between the frequency of development sites and overall tree removals.

Analysis of Geographic Trends:

Knowing where such impacts occur is equally as important as knowing what impacts have occurred. Viewing the data from a geographic perspective reveals some interesting trends. 
The map to the right shows all approved commercial development sites since  Jan 1, 2012. The most significant clustering is apparent in Austin's center. This region of Austin has seen the most frequent commercial development.
However, as before, tree removals do not follow the same pattern. Just as there is no strong correlation between frequency of development and total removals in a temporal sense, the same is true spatially.
The choropleth maps below contrast the relative frequency of development by council district with the relative proportion of total inches removed by council district. Whereas Austin's central region has seen the most frequent development, Austin's outskirts have seen the greatest volume of tree removals as a result of development. Tree removal has been particularly heavy in Austin's northwestern region. 
Removals, however, are only a part of the story.  The City's environmental ordinances typically require some proportion of removed inches to be replanted as mitigation. The exact proportion is contingent on a number of factors, such as health, species, and trunk size.   
Intuitively, it is reasonable to expect that the rate of mitigation (i.e. the ratio of inches replanted to inches removed) to be roughly even across the council districts. This is typically the case, with most council districts hovering close to the mean mitigation rate of 39%.

However,  the outliers yet again show some interesting characteristics.
One the one hand, District 4 has the fewest removals (7.5K inches),  yet it has the highest mitigation rate (51%). In contrast, District 6 has the most removals by far (133K inches), yet it has the lowest mitigation rate (29%)
A possible explanation is that District 4 is mostly urban, with a high proportion of impervious cover to green space. The remaining trees there tend to be protected size and native. Thus on those rare occasions when trees are removed, the average rate of mitigation tends to be higher. 
District 6, on the other hand, has several large undeveloped tracts and has many densely wooded areas. The trees there tend not be protected size, as the wooded areas are filled with young trees.  Also the trees there are more likely to be an undesirable species. Thus, tree removals are mitigated at a lesser rate on average.
Tree preservation is another important area of analysis. Preserved trees are any trees that were surveyed on a development site but were not removed. 
Ironically, District 6, the district with the most removals, appears to have a second highest number of trees preserved.  This fact is, again, likely due to the densely wooded areas present there. The abundance of trees in District 6 allows for both a high number of removed trees and a high number of preserved trees. 

Summary:

Exploring tree data from commercial development sites reveals some interesting and sometimes counter-intuitive trends.   There is, for instance, the lack of a strong correlation between the frequency of development sites and overall tree removals that expresses itself both temporally and spatially. There is also the fact that the area having the most removals also has the lowest average rate of mitigation, and, conversely, the area with the least removals also has the highest average rate of mitigation.  Lastly, there is the area having the most removals also having the second most trees preserved.  Overall, it appears that the story told by the data in this analysis is one of disproportion. While these facts are certainly puzzling, plausible explanations avail themselves. Nevertheless, this is just a small part of the story. So much remains to be uncovered, and further research is necessary.
Christopher Dolph
Chris is a Research Analyst with the Community Tree Preservation Division. He has been with the team since August of 2016. He has undergraduate degrees in Government and Philosophy, with a background in programming and computational data analysis, and is currently pursuing his Master's in Data Analytics. He enjoys spending time with his children, playing guitar, and nature.