IBM unveils a powerful new AI model, Prithvi-EO-2.0, designed to help communities worldwide combat climate change by predicting natural disasters, monitoring environmental changes, and optimizing renewable energy.
 |
| IBM's advanced AI model, Prithvi-EO-2.0, utilizes geospatial data to help communities better understand and prepare for the impacts of climate change, from predicting natural disasters to optimizing renewable energy. Symbolic Image |
Washington, USA --- January 31, 2025:
IBM has unveiled a powerful new geospatial artificial intelligence (AI) model called Prithvi-EO-2.0, designed to assist at-risk communities in understanding and preparing for the impacts of climate change.
This AI model, boasting 600 million parameters, is trained on high-resolution satellite imagery from NASA's HLS V2 product, enabling it to analyze data with 30-meter granularity. Prithvi-EO-2.0 can identify specific tree species and crops in satellite images, and even leverage drone footage to detect objects like cattle and solar panels.
Key to the model's functionality are the two types of data it's trained on: Harmonized Landsat Sentinel-2 (HLS) and Modern-Era Retrospective analysis for Research and Applications (MERRA-2).
HLS data empowers users to develop improved models for detecting wildfires, floods, and landslides. It also facilitates tasks like measuring canopy heights and tracking land use changes.
MERRA-2 data, on the other hand, is adept at tracking weather patterns, including hurricanes and turbulence. It can also be used for weather modeling and temperature mapping.
This new technology empowers countries that previously lacked access to such resources to plan for the consequences of rising global temperatures, safeguard their ecosystems, and issue timely warnings to populations in the path of climate-related disasters.
Prithvi-EO-2.0 boasts an enhanced Multi-Temporal Cloud Gap Imputation feature, allowing it to reconstruct areas obscured by clouds. This improved accuracy is crucial for tasks like crop yield estimation, agricultural planning, and environmental analysis.
IBM's commitment to environmental sustainability is exemplified by its 2023 partnership with the Kenyan government to support their reforestation initiatives. Prithvi-EO-2.0 has not only aided these efforts but has also helped secure funding for Kenya's reforestation projects and establish a flood warning system for vulnerable communities.
"The Global South is at the forefront of climate change," remarked Ambassador Philip Thigo, Kenya's special envoy on technology, according to TechRadar.
Geospatial AI has also proven valuable in detecting heat variations within African urban areas, enabling the modeling of health risks associated with heat islands and other climate-related issues.
IBM's long-term vision involves leveraging the knowledge gained from Prithvi-EO-2.0 to develop an open-source model called GridFM. This model is designed to manage power infrastructure and facilitate the transition to renewable energy.
GridFM will be capable of forecasting clean energy generation, monitoring power transmission, and aiding in predicting energy demands and costs.
"By 2050, power production will need to double in order to meet decarbonization goals," said François Mirallès, a researcher at Hydro Québec, as quoted by TechRadar. GridFM has the potential to significantly smoothen the clean energy transition.