Soil in the Atlantic Forest can store large amounts of carbon, helping to keep it out of the atmosphere

Art by Valentina Fraiz
Murilo Bomfim
The solution to reducing the carbon emissions that cause global warming may be closer than we think: right beneath our feet. In addition to supporting everything on the planet, soil acts as a kind of carbon guardian. In practice, the soil is quite greedy: once it takes up an element, it hardly ever gives it back up. This means soil can lock in carbon, keeping it out of the atmosphere and helping to lessen the greenhouse effect.
This idea shapes agronomist Alessandro Samuel-Rosa’s research. Growing up on a farm in the small town of Espumoso in Rio Grande do Sul, he has cultivated a relationship with the soil since childhood. He studied agricultural technology in high school and participated in an exchange program in Hawaii, where he learned about crop diversification and growing fruits and vegetables using agroecological methods—forward-thinking approaches for the time championed by his teachers. He even sought to use what he had learned on his family’s farm and introduce new crops, but his father refused to shift away from soybeans. Unable to put his teacher’s lessons into practice, he ultimately abandoned farming to become a teacher instead.
He started his undergraduate research on soils early in his studies at the Federal University of Santa Maria. He began by collecting field data and independently learning geostatistics to interpret it. Without realizing it, he was laying the groundwork for his entry into a new area of soil science, digital mapping, driven by the computational advances of the early 2000s.
Mapping soil is anything but simple. Unlike vegetation, which can be monitored with aerial imagery, soil lies concealed under plants, organic matter, buildings, and asphalt. Samples must be collected manually with tools such as shovels and hoes. Samuel-Rosa participated in soil-mapping projects during his master’s and doctoral studies. While doing postdoctoral work and prior to taking up his position as a professor at the Federal Technical University of Paraná (UTFPR), he created SoilData, Brazil’s largest soil data repository. SoilData is an open data platform that aggregates information from IBGE, Embrapa, universities, and private companies.
Samuel-Rosa’s work attracted the interest of MapBiomas Brasil, a collaborative network that creates annual maps of land cover and land use across the country. The network was founded in 2021 and partnered with Samuel-Rosa to estimate soil organic carbon (SOC). This partnership guided his research at the UTFPR Pedometrics Laboratory, led by Taciara Horst, toward this line of investigation.
Carbon mainly enters the soil through vegetation since plants are largely made of this element. As plant material falls to the ground, microorganisms break it down, converting some of the carbon into CO2 and releasing it into the atmosphere. Another part of the carbon, however, gets transported into the soil by ants, earthworms, and fungi.
Once in the soil, carbon forms strong chemical bonds with minerals, such as calcium and iron, which makes it less available to microorganisms. As a result, it is much less likely to be released into the atmosphere and contribute to the greenhouse effect. In other words, the most beneficial carbon is carbon that is bound to the soil.
Soil loses carbon primarily through deforestation and predatory farming. When natural plant cover is cleared, the soil becomes exposed to rain and sunlight, which activates microorganisms that release CO2. Predatory agricultural practices also expose the soil’s organic matter, breaking its bonds and releasing CO2. This process is particularly significant in Brazil, where deforestation and agriculture are the dominant sources of carbon emissions.
Soil is so resilient that even after major disturbances, SOC levels change relatively little over the years. At least, that is what the records from 1985 onward suggest. Samuel-Rosa’s current challenge is essentially a journey back in time. Using artificial intelligence and machine learning, he is trying to estimate how much SOC existed in Brazilian biomes before European colonization.
Soil is so resilient that even after major disturbances, SOC levels change relatively little over time. At least, that is what the records from 1985 onward, when this data collection began, suggest. Samuel-Rosa’s current challenge is essentially a journey back in time. Using artificial intelligence and machine learning, he is trying to estimate how much SOC existed in Brazilian biomes before European colonization.
Os serviços deste site podem conter links para outros sites ou serviços on-line que são operados e mantidos por terceiros e que não estão sob controle ou são mantidos pelo Instituto Serrapilheira.
Os artigos deste site podem incluir conteúdo incorporado de outros sites.
O site armazena as informações pessoais que os usuários registram em nosso site as quais só podem ser acessadas e editadas pelos administradores do site.
O usuário pode solicitar que apaguemos todos os dados pessoais que mantemos sobre ele. Isso não inclui nenhum dado que somos obrigados a manter para fins administrativos, legais ou de segurança.
Nenhum dado sobre os visitantes que se inscrevem no site é negociado pelo Instituto Serrapilheira, sob nenhuma circunstância.
O Instituto Serrapilheira pode decidir alterar a sua política de uso de dados a qualquer momento e por sua exclusiva deliberação.