{"id":27015,"date":"2026-02-12T12:13:53","date_gmt":"2026-02-12T17:13:53","guid":{"rendered":"https:\/\/www.clarku.edu\/news\/?post_type=story&p=27015"},"modified":"2026-02-13T11:11:51","modified_gmt":"2026-02-13T16:11:51","slug":"with-ais-assistance-researchers-maximize-clark-tool-to-reveal-how-our-planet-is-changing","status":"publish","type":"story","link":"https:\/\/www.clarku.edu\/news\/2026\/02\/12\/with-ais-assistance-researchers-maximize-clark-tool-to-reveal-how-our-planet-is-changing\/","title":{"rendered":"With AI\u2019s assistance, researchers maximize Clark tool to reveal how our planet is changing"},"content":{"rendered":"\n\n\n

Working with Professor Pontius, Ph.D. student Antonio Fonseca aims to help scientists understand the \u2018big picture\u2019 in decades of land-cover data<\/h3>\n\n\n\n

Growing up, attending college, and working in Bel\u00e9m, Brazil, Antonio Victor Galv\u00e3o da Fonseca observed firsthand how humans have changed the Amazon\u2019s landscape \u2014 clearing forests, building roads, expanding cropland, extracting minerals, and disrupting rivers.<\/p>\n\n\n\n

As a researcher at Imazon (Amazon Institute of People and the Environment), he produced maps and reports for policymakers on deforestation.<\/p>\n\n\n\n

Now a third-year Ph.D. student in geography<\/a> working with Professor Robert Gil Pontius Jr.<\/a> in Clark\u2019s School of Climate, Environment, and Society<\/a>, Fonseca is using artificial intelligence and cloud computing to improve a free, open-source research tool available on GitHub<\/a>. The tool allows users to process large datasets remotely and better analyze complex patterns of land change over time, in any region of the world.<\/p>\n\n\n\n

\"Antonio
Antonio Fonseca (photo courtesy of Imazon)<\/figcaption><\/figure>\n\n\n\n

\u201cWithout this tool, you would need a very powerful computer to analyze global land-cover data,\u201d Fonseca says. \u201cBut with the tool, anyone can process the data because all the processing happens in the cloud. If you have a regular computer, a web browser, and a Google account, you can use it.\u201d<\/p>\n\n\n\n

The tool relies on land-cover maps derived from satellite imagery of the Earth, captured over four decades. The maps classify the landscape into categories, such as forests, grassland, cropland, and water.<\/p>\n\n\n\n

Translating Pontius\u2019 mathematical equations into Python code, Fonseca has expanded the tool to work with maps containing over a dozen land-cover categories. The tool allows users to detect land change in maps and datasets from multiple time periods. For instance, it could illuminate how cropland has encroached on forests or how a river has shrunk over several decades or just two years.<\/p>\n\n\n\n

\u201cIn the past, we didn\u2019t have enough data,\u201d Fonseca says. \u201cNow we have too much data, and the limitation is understanding it. We need tools that help people see what is most important.\u201d<\/p>\n\n\n\n

With Clark AI Innovation Fund grant, researchers further develop tool for worldwide use<\/h4>\n\n\n\n
\n
\n

Although the tool deals solely with historical data, it could allow conservation and GIS scientists, policymakers, and regional land-use planners across the world to uncover the patterns which, combined with further research, could help them understand the impacts of climate change over time and make future projections, he explains.<\/p>\n\n\n\n

To support their research using AI to improve the land-cover data tool, Fonseca and Pontius were awarded a grant<\/a> from the Clark AI Innovation Fund, established with a $50,000 gift from the Yee Family, including Trustee Brian Yee \u201993. <\/p>\n<\/div>\n\n\n\n

\n