By Keith Allison
Ask the average person what lines the streets of New York City, and the likeliest response will be buildings. Nestled at the foot of the city’s brownstones, skyscrapers, and towers, however, are the city’s street trees. Although they can get lost in the shadows, New York City contains a lot of trees — some neighborhoods more than others — and those trees play a vital role in the urban ecosystem far beyond the obvious contribution of beautification. The distribution of trees, their health, their size, their species, all of these things play a crucial role in how the city works, so it’s important to know each of these things. Which is why, every ten years, the city’s Department of Parks and Recreation counts every street tree in the five boroughs (most recently in 2005) with the help of community groups and local volunteers. That’s over 33,000 blocks. Aside from the logistical and physical requirements of this TreesCount! project, there is the question of how the data is collected, how it’s presented, and what’s done with it. What does understanding the location and species of trees lining the city’s streets contribute to the understanding of how the city functions? What does it take to collate that much information into something decipherable? What can you do with that data?
Quite a lot, it turns out.
New York’s first tree census occurred in 1995 and collected data on the number, size, and species of street trees in the city. At the time, the data was largely collected by hand. Among the findings resulting from the effort were some 10,000 dead trees lining New York’s streets. The census enabled Parks NYC to create programs centered around pruning and replacing these trees. A decade later, in 2005, it was time to count again. This time, however, advances in GIS technology meant that not only could Parks NYC and its group of 1,100 volunteers collect information on number, size, and species; they could also take the data collected and plot it on an interactive map, then make that data available to researchers for use in their own studies.
Using a GeoSupport application developed by the New York City Department of Planning, the census data was mapped in ArcGIS. The result yielded a wealth of information on the impact of trees in an urban environment. This included, among other things, the impact of street trees on air quality, energy savings, carbon sequestration, and the reduction of stormwater runoff – as well as information on the impact of green spaces on the overall quality of life for the residents of New York’s five boroughs.
Mapping Research Potential
By the end of the 2005 census, researchers had used the data collected to determine that New York’s 592,130 street trees at the time generated nearly $122 million annually in benefits and savings for the city. The most exciting aspect of the census for researchers is the incredible number of ways in which it can be visualized and applied. This massive collection of data has been used to generate a number of maps that help researchers in many fields—from urban planning to environmental science to economics and sociology—study a system as complex as New York. It can be especially powerful when combined with other data sets collected by the city, such as those relating to economics.
Ben Wellington, a Ph.D. in Computer Science at NYU, used the data to investigate possible correlations between the density of trees in a given neighborhood and the economic well-being of that neighborhood. Writing for his blog I Quant NY: Quantitative Analysis of NYC Open Data, Wellington created street tree “heat maps” (visualizations that measure density and render a color-coded map) that displayed the density of trees in the city. He found that while there might be a correlation in some cases (wealthy neighborhoods like Brooklyn Heights predictably contain a lot of street trees while a neighborhood like East New York does not), many other factors impact the distribution of trees. These included the presence of large buildings (for all its economic strength, lower Manhattan’s Financial District contains almost no street trees) and the length and concentration of streets. Manhattan’s Greenwich Village, for example, is a neighborhood covered by trees, but the number of trees is surprisingly small. It’s the number of streets and the relative shortness of the blocks that creates the density of green.
The discovery that, between 1995 and 2005, the city had increased its number of street trees by 19% (meaning that, as of 2005, street trees accounted for about a quarter of New York’s overall “urban canopy” — the collection of foliage that calls New York City home) led to then mayor Michael Bloomberg launching MillionTrees NYC. How close the initiative has brought the city to a million street trees will be revealed during the 2015 census, which, among other things, will provide the city with a chance to analyze the impact of two major hurricanes (Irene in 2011 and Sandy in 2012) on the street tree population as well as the potential impact of the trees on neighborhoods during the hurricane (where did they protect? where did they do damage?).
The data also allows researchers to analyse the types of trees present in New York’s urban ecosystem and study the unique benefits (or lack thereof) from specific species. The 2005 census found 168 species make up the city’s street tree population, and some 15% of those are one species: the London Planetree. Why so many of that particular species? No scientific reason, it turns out. City planner Robert Moses just liked them. As a result of the census, and knowing that too great a concentration of any one species makes for too delicate an ecosystem, Parks NY has begun concentrating on increasing biodiversity. Other species include the incredibly resilient Ginkgo tree, which make up about 3% of the total and the pungent seed pods of which cover city streets in places like Washington Square Park.
Knowing the distribution of tree species helps Parks NYC increase biodiversity and predict how certain trees might negatively impact a neighborhood. Staten Island, for example, has a lot of Callery Pear trees, a species that loses limbs easily, thus making them more likely to damage property, and has proven unwelcoming to birds and other animals that promote a healthy ecosystem. They also smell awful.
One of the most talked about uses of the 2005 census data came courtesy of Brooklyn web designer Jill Hubley, who fed the data (downloadable as a CSV file for each borough) into CartoDB and used it to generate an interactive online map that enables people to filter the map by species. Creating the map took some additional labor, mostly centered around the fact that the census data refers to each tree species by a code. That code has to be translated into the actual name of the species for it to be meaningful to researchers. As Hubley wrote, “The original data only included these tree codes, e.g., PLAC for London Plane. The first two digits of the code are the beginning of the tree’s genus (in this example, Platanus) and the last two digits are the beginning of the particular epithet for the species (in this example case, acerfolia.)”
The census data doesn’t even have to be used to analyze the trees themselves. NYU professor Joanna Klukowska with the Department of Computer Science at Courant Institute of Mathematical Sciences, needed a massive, varied data set for her students to use as part of a project to learn coding. “For the Data Structures class, I am always on the lookout for large data sets for the students to work with,” Dr. Klukowska explained. “The advantages of certain data organization approaches only becomes noticeable when you deal with hundreds of thousands of elements rather than two or three hundred. I also wanted to work with a real data set for two reasons: students get a sense of doing something that is applicable, and real data is not perfect so students need to deal with possible errors, missing fields, etc.” Her students discovered, like Jill Hubley, that the use of codes for tree species might help trim the size of the file, but it makes it difficult to work with the data.
While looking through the massive reservoir of data sets at NYC Open Data, she came across the tree census information and determined that it would be perfect. It had multiple potential ways of being organized (by zip code, tree type, diameter, etc.) and presented data sets of different sizes depending on the borough. For the course, the census data served a strictly technical purpose, but students were thrilled to learn such data existed. “We were not looking for any environmental conclusions. But students did comment about some surprises that they observed, like the huge differences in the number of entries in files for different boroughs and which zip codes turned out to be the greenest.”
The 2015 census benefits from even greater developments in GIS and mapping, allowing the city to pinpoint the location and impact of trees with an incredible degree of accuracy. This time around, Parks NYC is using an application called TreeKit, part of an initiative sponsored by the Open Space Institute through its Citizen Action Program, developed expressly for the task of collecting and mapping this sort of data. TreeKit equips volunteers with a mobile tablet and a guide to identifying tree species – just in case you can’t walk up and instantly tell a London Planetree from a Silver Maple. Parks NYC tested TreeKit by using it to map some 10,000 street trees in western Queens. In 2013, TreeKit worked with partners OpenPlans and CartoDB to create a mobile data entry tool that greatly increases both the ease of entering information and the accuracy of spatial data. But not everything is digital technology. Volunteers still use old-fashioned tape measures and measuring wheels to collect data like tree circumference and distance from one tree to the next. Parks NYC also created a visual guide to tree species that can be printed, and iPhone users can install LeafSnap, an app that identifies tree species based on photos you snap of the leaves.
As in previous years, all of that data will be publicly accessible via Open Data New York for researchers to download and use. Data sets for the 1995 census remain available, as do each of the sets for 2005: Manhattan, Brooklyn, Queens, The Bronx, and Staten Island. Those interested in volunteering to help survey the trees can sign up with Parks NYC. The census is expected to run throughout the fall of 2015.