Researchers create “pruning software” to reduce risk of urban tree falls

The technique uses laser equipment to scan and create three-dimensional images; the algorithm optimizes the cut, seeking to maintain the balance and health of the tree.

By Luciana Constantino  |  Agência FAPESP – Researchers at the University of São Paulo (USP) in Brazil have developed a software that accurately guides the pruning of trees in urban areas, reducing the risk of them falling. With more frequent and extreme weather events, storms in large cities have felled a greater number of trees, causing economic losses and loss of life.

In São Paulo alone, after being subject to heavy rains, the city government counted 330 fallen trees in a single day (March 12th) – one of which hit a car, killing the driver. It is estimated that around 2,000 urban trees fall every year in the city, not including public parks and protected areas.

According to the scientists, the first step is to scan the trees with laser equipment to take three-dimensional images at various points, which are then placed in a computer model. To do the modeling, the group developed a combination of software and came up with a “pruning algorithm” to keep the tree balanced.

The work brings together biology, mathematics and mechatronics and is being developed by researchers from the Institute for Advanced Studies (IEA-USP), the Institute of Biosciences (IB-USP) and the Engineering School (POLI-USP) within the framework of the Research Center for Greenhouse Gas Innovation (RCGI), an Engineering Research Center (ERC) created by FAPESP and Shell, with the support of several other companies, aimed at advanced studies in bioenergy, agriculture and environmental sciences to mitigate the effects of climate change.

“Trees, like other forms in nature, are optimized structures that use only the essential material in their physical configuration. If you prune the wrong way, you can weaken that structure. With the software we developed, we were able to assess the deformations, even the deflection of trees due to wind. With this, it’s possible to have a diagnosis and guide pruning more precisely,” Emílio Carlos Nelli Silva, professor at the Department of Mechatronic Engineering and Mechanical Systems at POLI-USP and deputy scientific director of the RCGI, told Agência FAPESP.

In the first analyses, carried out on the USP campus in São Paulo, the researchers pruned up to 20% of the crown using the algorithm and managed to maintain the balance of the tree. In another test, with a different type of cut, there was an increase in the number of weak points, with a greater chance of falling.

Normally, pruning in urban areas is carried out as part of a municipality’s environmental management plan to make the structure of the plant compatible with human interaction, with a focus on the health of the tree, maintaining its resilience and balance. The definition of how each tree will be pruned is made by the technicians or professionals in charge of the service, within a limit of 30% of the crown.

“Humans don’t have the ability to do this pruning as precisely as we can using algorithmic scanning. We haven’t yet calculated the percentage of trees that the software can prevent from falling with more precise pruning, but we will soon,” says Marcos Buckeridge, professor in the Department of Botany at IB-USP and deputy director of the IEA.

He adds: “This tool comes from a sequence of work that now also includes the engineering team. The next step is to work with meteorology to study the speed and direction of winds during storms. This will help us detect the weakest side of the crown or whether the tree is in a wind channel. That way we can improve pruning and avoid planting in windy areas.”

Research published last year showed that drastic pruning, along with the condition of the wood and root strangulation by sidewalks, can be used as predictors of urban tree toppling (read more at: agencia.fapesp.br/51060).

Among the main factors in urban areas that influence tree fall are the height of the surrounding buildings, the age of the neighborhood, the width of the sidewalk, and the height of the tree. As they verticalize, they face unfavorable conditions in so-called “urban canyons,” i.e., continuous rows of tall buildings that alter local wind speed, pollution dispersion, and lighting and microclimate patterns, contributing to premature decline (read more at: agencia.fapesp.br/39670). 

“Newton’s apple”

Nelli Silva says that the idea for the software came about during a lunch with Buckeridge at a restaurant in São Paulo, where a century-old Figueira-de-bengala (Ficus benghalensis) tree was planted, probably in the 1890s. It is a member of the Ficus genus and stands about 6 meters tall, with a crown diameter of about 46 meters.

“I knew that Buckeridge was working on this line of study, and I’ve always been curious about trees with complex structures that are difficult to model on a computer. So we thought: why not use an algorithm to restore the optimization of these structures through repodding? That was the inspiration for the research,” recalls the professor, who specializes in computational mechanics.

Nelli Silva explains that nature often inspires structural optimization studies, and fruits such as apples can serve as a model for analysis. “The apple shape is a structure that has the lowest mechanical stress subject to its own weight. It has a mass resting on one point, which is the stalk. We started from the idea that if it’s possible to analyze the optimized structure of the apple, it’d also be possible with the tree.”

In science, a famous story about the fruit that continues to generate research and questions today is that of the English physicist Isaac Newton, who is said to have conceived the Law of Universal Gravitation while observing an apple falling from its tree. The law states that the force of attraction between two bodies must be proportional to the product of their masses divided by the square of the distance between them.

Next steps

By modeling the images obtained by the laser equipment, it is possible to arrive at a kind of “architecture” of the tree, a step that involved the collaboration of USP professor Marcelo Zuffo, one of the researchers of the project “CID-SP EMERGENCIAS: Public Health Data Intelligence Center”, supported by FAPESP through the program Science Centers for Development (SCD).

In this first stage, the analysis was carried out on the crown. The group now intends to incorporate information from the roots. “The methodology opens up the prospect of doing new computer simulations, customized for each species and with other types of data. The method is scalable enough to map the trees of an entire city,” says Nelli Silva.

According to Buckeridge, some city governments, including that of São Paulo, have already approached the group about using the software. “We’re connecting science with technology development to address climate change. We want to increase the speed of analysis and reduce the cost of the software. This paves the way for startups that want to help with the service. The more trees we have, the better the resilience of cities to climate change,” adds the deputy director of the IEA, which is part of the National Institute of Science and Technology (INCT) for Bioethanol, funded by FAPESP and the National Council for Scientific and Technological Development (CNPq).

When contacted for this report, São Paulo’s Department of Greenery and the Environment (SVMA) said it was “in talks to get to know the software and assess its applicability in the city’s tree management services,” as part of its commitment to adopting technologies that improve environmental management.

The department says it intends to analyze how the tool can contribute to preventative pruning and reduce the risk of trees falling, especially during storms.