All News
February 27, 2025

Sounding a Warning for Urban Biodiversity

By Jared Brey
Map showing bands for biodiversity loss risk, climate impact risk, and development suitability
Hotspot Stoplight uses a series of scripts to connect geospatial data to machine-learning models that classify land types and predict changes in land cover. ​This map shows our risk and suitability indices for Costa Rica in a disaggregated way, with biodiversity loss risk in red, climate impact risk in purple and urban development suitability in black. Each of these indices varies across space, which is shown in this visualization through the varying opacity of each color.
View Slideshow
Close
Previous
Next
Map showing bands for biodiversity loss risk, climate impact risk, and development suitability
Map showing Low-risk development in central San José with high-risk encroaching to the south
Hotspot map
Map of San Jose region showing different types of sites visited, urban and rural
Hotspot Stoplight uses a series of scripts to connect geospatial data to machine-learning models that classify land types and predict changes in land cover. ​This map shows our risk and suitability indices for Costa Rica in a disaggregated way, with biodiversity loss risk in red, climate impact risk in purple and urban development suitability in black. Each of these indices varies across space, which is shown in this visualization through the varying opacity of each color.
This map shows the suitability of urban development, based on predicted probability of urban land cover minus biodiversity loss risk, climate impact risk, and biodiversity intactness. Different segments of the spectrum represent potential suitability for different urban strategies.
The team used a clustering-based analysis on the components of each index to pinpoint zones of potential suitability for different urban strategies. The ground-truthing process involved visiting dozens of sites identified by this process to investigate the accuracy the approach.
The Weitzman team's ground-truthing of the Hotspot Spotlight tool involved visits various sites across the San José.
Areas
Tags
Media Contact

Michael Grant
mrgrant@design.upenn.edu
215.898.2539

Last year, the designers Matthijs Bouw and Oliver Atwood (MLA’24, MUSA'24) were in San José, Costa Rica, to test out a tool they were developing to protect biodiversity and reduce climate risk in places where biodiversity is threatened by uncontrolled urbanization.

The tool, called Hotspot Stoplight, is meant to help cities understand where urban growth is risky and threatening sensitive natural habitats. Bouw, a professor of practice in landscape architecture and architecture at the Weitzman School, and Atwood, a former student who is now a designer at Bouw’s firm, One Architecture & Urbanism, were in Costa Rica to “ground-truth” its effectiveness. They had driven to an area on the edge of the city where their maps had shown a high likelihood of biodiversity loss. What they found there was a forest being actively clear cut—“potentially clandestine logging,” Atwood says.

The Hotspot Stoplight tool, being developed in collaboration with UN-Habitat, the United Nations program charged with promoting socially and environmentally sustainable towns and cities, grew out of a broader research project on urban biodiversity hotspots that Bouw began in 2018 with Professor Emeritus of Landscape Architecture Richard Weller. It was rooted in research into “Hotspot Cities” that Weller began in 2013, investigating land-use patterns in the world’s most biodiverse places. Weller is also the author of the 2024 book To the Ends of the Earth: A Grand Tour for the 21st Century, for which Atwood provided the illustrations, exploring human imprints on the global landscape. Bouw and Atwood are hoping the new tool will help raise awareness of the links and conflicts between urban development and sensitive wildlife areas. They’re also hoping it can help local officials determine appropriate places for new development.

Hilly, forested terrain showing swaths of lands that have been clear cut

This image shows an area that the Hotspot Spotlight algorithm predicted was a forest with high biodiversity loss risk. Upon arrival there, it became clear that that risk had been realized—a logging operation was actively underway. (Photo Oliver Atwood)

“If you talk to mayors locally, they are not interested so much in where you don’t want to develop if you don’t also give them the areas where they should develop,” Bouw says. “Everyone wants the tax revenue and the growth and wants to find ways to accommodate people in the metropolitan areas.”

Bouw and Atwood created the technical components of the tool using publicly available global datasets. They used a series of scripts to connect geospatial data to machine-learning models that classify land types and predict changes in land cover. It’s calibrated to a 30-meter resolution, and designed to pinpoint areas of potential biodiversity loss.

But to test the tool’s accuracy, they needed to visit some sites in person. They chose Costa Rica because its elected officials are uniquely attuned to the environmental and cultural benefits of natural biodiversity. Lucía Artavia (MCP’16), a planning and policy advisor who’d worked with Weller on several projects she was a student at Weitzman, helped introduced Bouw and Atwood to local officials. She also helped them review their GIS analysis and plan visits to specific sites. A native of Costa Rica, Artavia had returned after graduation and worked with government officials in the district of Curridabat promoting pollinator habitat in urban spaces.

Matthijs Bouw
"There’s a receptiveness to these issues," says Matthijs Bouw, a professor of practice in landscape architecture and architecture, of his team's work with local leaders to protect biodiversity in Costa Rica.

“Local governments have very scarce resources, financially and technically, and having this input, this analysis, and these kinds of technical resources is quite useful to them,” says Artavia. “We brand ourselves as a very environmentally friendly country, and we do have beautiful natural reserves, but when it comes to our cities, they’re completely hostile to nature.”

The stakes of the trip to San José weren’t limited to the accuracy of the Hotspot Spotlight’s mapping capabilities, says Bouw. The tool needs to promote productive conversations among local leaders in the cities where it’s being tested—conversations about where to promote growth, where to restrict it, and where to carefully protect existing habitats. The tool also needs to guide officials in determining where to prioritize investments in nature-based infrastructure.

“Costa Rica is just a place that’s very well known for its biodiversity. There are a lot of good scientists there. There’s a receptiveness to these issues,” Bouw says.

Atwood is still calibrating the tool for greater precision. For example, future versions should be able to distinguish between specific types of agricultural land uses, he says. But international advocacy groups have already shown interest in using Hotspot Spotlight to help cities create better practices around development and conservation. And the tool can be optimized for specific places to be more useful to local leaders working with minimal resources.

“We want to better understand biodiversity loss and climate risks in cities around the world, and specifically in cities that are not the big metropolises,” Bouw says. “These cities don’t have the capacity to understand what is coming at them.”

More News