Jump to content

Expanded AI Model with Global Data Enhances Earth Science Applications 


Recommended Posts

  • Publishers
Posted

4 min read

Expanded AI Model with Global Data Enhances Earth Science Applications 

A false-color image of the East Peak fire burning in southern Colorado near Trinidad, captured by the Operational Land Imager on the Landsat 8 satellite. Dark green forests and light green grasslands cover most of the image, but a red patch in the middle represents a burn scar, and some orange spots around it represent actively burning areas.
On June 22, 2013, the Operational Land Imager (OLI) on Landsat 8 captured this false-color image of the East Peak fire burning in southern Colorado near Trinidad. Burned areas appear dark red, while actively burning areas look orange. Dark green areas are forests; light green areas are grasslands. Data from Landsat 8 were used to train the Prithvi artificial intelligence model, which can help detect burn scars.
NASA Earth Observatory

NASA, IBM, and Forschungszentrum Jülich have released an expanded version of the open-source Prithvi Geospatial artificial intelligence (AI) foundation model to support a broader range of geographical applications. Now, with the inclusion of global data, the foundation model can support tracking changes in land use, monitoring disasters, and predicting crop yields worldwide. 

The Prithvi Geospatial foundation model, first released in August 2023 by NASA and IBM, is pre-trained on NASA’s Harmonized Landsat and Sentinel-2 (HLS) dataset and learns by filling in masked information. The model is available on Hugging Face, a data science platform where machine learning developers openly build, train, deploy, and share models. Because NASA releases data, products, and research in the open, businesses and commercial entities can take these models and transform them into marketable products and services that generate economic value. 

“We’re excited about the downstream applications that are made possible with the addition of global HLS data to the Prithvi Geospatial foundation model. We’ve embedded NASA’s scientific expertise directly into these foundation models, enabling them to quickly translate petabytes of data into actionable insights,” said Kevin Murphy, NASA chief science data officer. “It’s like having a powerful assistant that leverages NASA’s knowledge to help make faster, more informed decisions, leading to economic and societal benefits.”

AI foundation models are pre-trained on large datasets with self-supervised learning techniques, providing flexible base models that can be fine-tuned for domain-specific downstream tasks.

Three images show the process of crop classification with NASA and IBM’s open-source Prithvi Geospatial artificial intelligence model. The first shows a true color composite image of a cropland area. The second shows a Ground Truth Mask of the types of land cover in the image — natural vegetation is colored red, forest is orange, corn is yellow, soybeans are light green, wetlands are mid-green, developed or barren land is darker green, open water is light blue, winter wheat is mid-blue, alfalfa is dark blue, fallow/idle cropland is dark purple, cotton is pink, and sorghum is dark pink. The third image shows a Predicted Mask that closely matches the Ground Truth Mask.
Crop classification prediction generated by NASA and IBM’s open-source Prithvi Geospatial artificial intelligence model.

Focusing on diverse land use and ecosystems, researchers selected HLS satellite images that represented various landscapes while avoiding lower-quality data caused by clouds or gaps. Urban areas were emphasized to ensure better coverage, and strict quality controls were applied to create a large, well-balanced dataset. The final dataset is significantly larger than previous versions, offering improved global representation and reliability for environmental analysis. These methods created a robust and representative dataset, ideal for reliable model training and analysis. 

The Prithvi Geospatial foundation model has already proven valuable in several applications, including post-disaster flood mapping and detecting burn scars caused by fires.

One application, the Multi-Temporal Cloud Gap Imputation, leverages the foundation model to reconstruct the gaps in satellite imagery caused by cloud cover, enabling a clearer view of Earth’s surface over time. This approach supports a variety of applications, including environmental monitoring and agricultural planning.  

Another application, Multi-Temporal Crop Segmentation, uses satellite imagery to classify and map different crop types and land cover across the United States. By analyzing time-sequenced data and layering U.S. Department of Agriculture’s Crop Data, Prithvi Geospatial can accurately identify crop patterns, which in turn could improve agricultural monitoring and resource management on a large scale. 

The flood mapping dataset can classify flood water and permanent water across diverse biomes and ecosystems, supporting flood management by training models to detect surface water. 

Wildfire scar mapping combines satellite imagery with wildfire data to capture detailed views of wildfire scars shortly after fires occurred. This approach provides valuable data for training models to map fire-affected areas, aiding in wildfire management and recovery efforts.

Three images show the process of burn scar mapping with NASA and IBM’s open-source Prithvi Geospatial artificial intelligence model.The first shows a true color composite satellite image, which contains shades of green and purple. The second shows a Ground Truth Mask, which shows the true extent of a burn scar on the land by blacking out the area around the burn scar. The third shows a Predicted Mask, which almost exactly matches the Ground Truth Mask.
Burn scar mapping generated by NASA and IBM’s open-source Prithvi Geospatial artificial intelligence model.

This model has also been tested with additional downstream applications including estimation of gross primary productivity, above ground biomass estimation, landslide detection, and burn intensity estimations. 

“The updates to this Prithvi Geospatial model have been driven by valuable feedback from users of the initial version,” said Rahul Ramachandran, AI foundation model for science lead and senior data science strategist at NASA’s Marshall Space Flight Center in Huntsville, Alabama. “This enhanced model has also undergone rigorous testing across a broader range of downstream use cases, ensuring improved versatility and performance, resulting in a version of the model that will empower diverse environmental monitoring applications, delivering significant societal benefits.”

The Prithvi Geospatial Foundation Model was developed as part of an initiative of NASA’s Office of the Chief Science Data Officer to unlock the value of NASA’s vast collection of science data using AI. NASA’s Interagency Implementation and Advanced Concepts Team (IMPACT), based at Marshall, IBM Research, and the Jülich Supercomputing Centre, Forschungszentrum, Jülich, designed the foundation model on the supercomputer Jülich Wizard for European Leadership Science (JUWELS), operated by Jülich Supercomputing Centre. This collaboration was facilitated by IEEE Geoscience and Remote Sensing Society.  

For more information about NASA’s strategy of developing foundation models for science, visit https://science.nasa.gov/artificial-intelligence-science.

Share

Details

Last Updated
Dec 04, 2024

Related Terms

View the full article

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Similar Topics

    • By NASA
      Scientists find that cometary dust affects interpretation of spacecraft measurements, reopening the case for comets like 67P as potential sources of water for early Earth. 
      Researchers have found that water on Comet 67P/Churyumov–Gerasimenko has a similar molecular signature to the water in Earth’s oceans. Contradicting some recent results, this finding reopens the case that Jupiter-family comets like 67P could have helped deliver water to Earth.  
      Water was essential for life to form and flourish on Earth and it remains central for Earth life today. While some water likely existed in the gas and dust from which our planet materialized around 4.6 billion years ago, much of the water would have vaporized because Earth formed close to the Sun’s intense heat. How Earth ultimately became rich in liquid water has remained a source of debate for scientists.
      Research has shown that some of Earth’s water originated through vapor vented from volcanoes; that vapor condensed and rained down on the oceans. But scientists have found evidence that a substantial portion of our oceans came from the ice and minerals on asteroids, and possibly comets, that crashed into Earth. A wave of comet and asteroid collisions with the solar system’s inner planets 4 billion years ago would have made this possible.   
      This image, taken by ESA’s Rosetta navigation camera, was taken from a about 53 miles from the center of Comet 67P/Churyumov-Gerasimenko on March 14, 2015. The image resolution is 24 feet per pixel and is cropped and processed to bring out the details of the comet’s activity. ESA/Rosetta/NAVCAM While the case connecting asteroid water to Earth’s is strong, the role of comets has puzzled scientists. Several measurements of Jupiter-family comets — which contain primitive material from the early solar system and are thought to have formed beyond the orbit of Saturn — showed a strong link between their water and Earth’s. This link was based on a key molecular signature scientists use to trace the origin of water across the solar system.
      This signature is the ratio of deuterium (D) to regular hydrogen (H) in the water of any object, and it gives scientists clues about where that object formed. Deuterium is a rare, heavier type — or isotope — of hydrogen. When compared to Earth’s water, this hydrogen ratio in comets and asteroids can reveal whether there’s a connection.  
      Because water with deuterium is more likely to form in cold environments, there’s a higher concentration of the isotope on objects that formed far from the Sun, such as comets, than in objects that formed closer to the Sun, like asteroids. 
      Measurements within the last couple of decades of deuterium in the water vapor of several other Jupiter-family comets showed similar levels to Earth’s water. 
      “It was really starting to look like these comets played a major role in delivering water to Earth,” said Kathleen Mandt, planetary scientist at NASA’s Goddard Space Flight Center in Greenbelt, Maryland. Mandt led the research, published in Science Advances on Nov. 13, that revises the abundance of deuterium in 67P. 

      About Kathleen Mandt

      But in 2014, ESA’s (European Space Agency) Rosetta mission to 67P challenged the idea that Jupiter-family comets helped fill Earth’s water reservoir. Scientists who analyzed Rosetta’s water measurements found the highest concentration of deuterium of any comet, and about three times more deuterium than there is in Earth’s oceans, which have about 1 deuterium atom for every 6,420 hydrogen atoms.  
      “It was a big surprise and it made us rethink everything,” Mandt said.  
      Mandt’s team decided to use an advanced statistical-computation technique to automate the laborious process of isolating deuterium-rich  water in more than 16,000 Rosetta measurements. Rosetta made these measurements in the “coma” of gas and dust surrounding 67P. Mandt’s team, which included Rosetta scientists, was the first to analyze all of the European mission’s water measurements spanning the entire mission. 
      The researchers wanted to understand what physical processes caused the variability in the hydrogen isotope ratios measured at comets. Lab studies and comet observations showed that cometary dust could affect the readings of the hydrogen ratio that scientists detect in comet vapor, which could change our understanding of where comet water comes from and how it compares to Earth’s water. 
      What are comets made of? It’s one of the questions ESA’s Rosetta mission to comet 67P/Churyumov-Gerasimenko wanted to answer. “So I was just curious if we could find evidence for that happening at 67P,” Mandt said. “And this is just one of those very rare cases where you propose a hypothesis and actually find it happening.” 
      Indeed, Mandt’s team found a clear connection between deuterium measurements in the coma of 67P and the amount of dust around the Rosetta spacecraft, showing that the measurements taken near the spacecraft in some parts of the coma may not be representative of the composition of a comet’s body.  
      As a comet moves in its orbit closer to the Sun, its surface warms up, causing gas to release from the surface, including dust with bits of water ice on it. Water with deuterium sticks to dust grains more readily than regular water does, research suggests. When the ice on these dust grains is released into the coma, this effect could make the comet appear to have more deuterium than it has.  
      Mandt and her team reported that by the time dust gets to the outer part of the coma, at least 75 miles from the comet body, it is dried out. With the deuterium-rich water gone, a spacecraft can accurately measure the amount of deuterium coming from the comet body.
      This finding, the paper authors say, has big implications not only for understanding comets’ role in delivering Earth’s water, but also for understanding comet observations that provide insight into the formation of the early solar system.  
      “This means there is a great opportunity to revisit our past observations and prepare for future ones so we can better account for the dust effects,” Mandt said. 
      By Lonnie Shekhtman
      NASA’s Goddard Space Flight Center, Greenbelt, Md.
      Explore More
      9 min read Towards Autonomous Surface Missions on Ocean Worlds


      Article


      31 mins ago
      1 min read Coming Spring 2025: Planetary Defenders Documentary
      ow would humanity respond if we discovered an asteroid headed for Earth? NASA’s Planetary Defenders…


      Article


      52 mins ago
      5 min read What’s Up: December 2024 Skywatching Tips from NASA


      Article


      1 day ago
      Share








      Details
      Last Updated Dec 03, 2024 Editor Lonnie Shekhtman Contact Lonnie Shekhtman lonnie.shekhtman@nasa.gov Location Goddard Space Flight Center Related Terms
      Comets Goddard Space Flight Center Planetary Science Planetary Science Division Rosetta Science Mission Directorate The Solar System View the full article
    • By European Space Agency
      Researchers have achieved a breakthrough in bringing 5G technology to space, marking a pivotal moment in telecommunications history. The 5G Infrastructure Study (5G-IS), funded by ESA’s Connectivity and Secure Communications directorate, has mapped out how satellites can deliver reliable 5G connectivity worldwide.
      View the full article
    • By European Space Agency
      Image: A colourful patchwork of agricultural fields is pictured in this radar image captured by Copernicus Sentinel-1 over southeastern Romania. View the full article
    • By European Space Agency
      As the launch of the Sentinel-1C satellite approaches, we reflect on some of the many ways the Copernicus Sentinel-1 mission has given us remarkable radar insights into our planet over the years.
      View the full article
    • By NASA
      4 min read
      NASA AI, Open Science Advance Natural Disaster Research and Recovery
      Hurricane Ida is pictured as a category 2 storm from the International Space Station as it orbited 264 miles above the Gulf of Mexico. In the foreground is the Canadarm2 robotic arm with Dextre, the fine-tuned robotic hand, attached. NASA By Lauren Perkins
      When you think of NASA, disasters such as hurricanes may not be the first thing to come to mind, but several NASA programs are building tools and advancing science to help communities make more informed decisions for disaster planning. 
      Empowered by NASA’s commitment to open science, the NASA Disasters Program supports disaster risk reduction, response, and recovery. A core element of the Disasters Program is providing trusted, timely, and actionable data to aid organizations actively responding to disasters.  
      Hurricane Ida made landfall in Louisiana Aug. 21, 2021, as a category 4 hurricane, one of the deadliest and most destructive hurricanes in the continental United States on record. The effects of the storm were widespread, causing devastating damage and affecting the lives of millions of people. 
      During Hurricane Ida, while first responders and other organizations addressed the storm’s impacts from the ground, the NASA Disasters program was able to provide a multitude of remotely sensed products. Some of the products and models included information on changes in soil moisture, changes in vegetation, precipitation accumulations, flood detection, and nighttime lights to help identify areas of power outages.
      Image Before/After The NASA team shared the data with its partners on the NASA Disasters Mapping Portal and began participating in cross-agency coordination calls to determine how to further aid response efforts. To further connect and collaborate using open science efforts, NASA Disasters overlaid publicly uploaded photos on their Damage Proxy Maps to provide situational awareness of on-the-ground conditions before, during, and after the storm.  
      Immediate post-storm response is critical to saving lives; just as making informed, long- term response decisions are critical to providing equitable recovery solutions for all. One example of how this data can be used is blue tarp detection in the aftermath of Hurricane Ida.
      Using artificial intelligence (AI) with NASA satellite images, the Interagency Implementation and Advanced Concepts Team (IMPACT), based at NASA’s Marshall Space Flight Center in Huntsville, Alabama, conducted a study to detect the number of blue tarps on rooftops in the aftermath of hurricanes, such as Ida, as a way of characterizing the severity of damage in local communities.
      An aerial photograph shows damaged roofs from Hurricane Maria in 2017 in Barrio Obrero, Puerto Rico. In the wake of the hurricane, the Federal Emergency Management Agency (FEMA) and United States Army Corps of Engineers distributed 126,000 blue tarps and nearly 60,000 temporary blue roofs to people awaiting repairs on damaged homes. NASA While disasters cannot be avoided altogether, timely and accessible information helps communities worldwide reduce risk, improve response, hasten recovery, and build disaster resilience.  
      Through an initiative led by NASA’s Office of the Chief Science Data Officer, NASA and IBM are developing five open-source artificial intelligence foundation models trained on NASA’s expansive satellite repositories. This effort will help make NASA’s vast, ever-growing amounts of data more accessible and usable. Leveraging NASA’s AI expertise allows users to make faster, more informed decisions. User applications of the Prithvi Earth Foundation Models could range from identifying flood risks and predicting crop yields to forecasting long range atmospheric weather patterns.
      “NASA is dedicated to ensuring that our scientific data are accessible and beneficial to all. Our AI foundation models are scientifically validated and adaptable to new data, designed to maximize efficiency and lower technical barriers. This ensures that even in the face of challenging disasters, response teams can be swift and effective,” said Kevin Murphy, NASA’s chief science data officer. “Through these efforts, we’re not only advancing scientific frontiers, but also delivering tangible societal benefits, providing data that can safeguard lives and improve resilience against future threats.” 
      Hear directly from some of the data scientists building these AI models, the NASA disaster response team, as well as hurricane hunters that fly directly into these devastating storms on NASA’s Curious Universe podcast. 
      Learn more about NASA’s AI for Science models at https://science.nasa.gov/artificial-intelligence-science/.
      Share








      Details
      Last Updated Nov 26, 2024 Related Terms
      Earth Natural Disasters Open Science Explore More
      5 min read NASA Data Reveals Role of Green Spaces in Cooling Cities


      Article


      3 hours ago
      5 min read 5 Surprising NASA Heliophysics Discoveries Not Related to the Sun


      Article


      6 days ago
      14 min read NASA’s Brad Doorn Brings Farm Belt Wisdom to Space-Age Agriculture
      From his South Dakota roots to leading NASA’s agricultural program, Brad Doorn’s mission has remained…


      Article


      6 days ago
      Keep Exploring Discover More Topics From NASA
      Missions



      Humans in Space



      Climate Change



      Solar System


      View the full article
  • Check out these Videos

×
×
  • Create New...