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
      3 min read
      Preparations for Next Moonwalk Simulations Underway (and Underwater)
      When most people think of NASA, they picture rockets, astronauts, and the Moon. But behind the scenes, a group of inventors is quietly rewriting the rules of what’s possible — on Earth, in orbit, and beyond. Their groundbreaking inventions eventually become technology available for industry, helping to shape new products and services that improve life around the globe. For their contributions to NASA technology, we welcome four new inductees into the 2024-2025 NASA Inventors Hall of Fame

      A robot for space and the workplace

      Myron (Ron) Diftler led the team behind Robonaut 2 (R2), a humanoid robot developed with General Motors. The goal was to create a robot that could help humans both in space and on the factory floor. The R2 robot became the first humanoid robot in space aboard the International Space Station, and part of its technology was licensed for use on Earth, leading to a grip-strengthening robotic glove to help humans with strenuous, repetitive tasks. From factories to space exploration, Diftler’s work has real-world impact. 

      Some of the toughest electronic chips on and off Earth

      Technology developed to one day explore the surface of Venus has to be tough enough to survive the planet where temperatures hit 860°F and the atmosphere is akin to battery acid. Philip Neudeck’s silicon carbide integrated circuits don’t just work — they ran for over 60 days in simulated Venus-like conditions. On Earth, these chips can boost efficiency in wireless communication systems, help make drilling for oil safer, and enable more practical electric vehicles. 
      From developing harder chip materials to unlocking new planetary missions, Neudeck is proving that the future of electronics isn’t just about speed — it’s about survival.

      Hydrogen sensors that could go the distance on other worlds

      Gary Hunter helped develop a hydrogen sensor so advanced it’s being considered for a future mission to Titan, Saturn’s icy moon. These and a range of other sensors he’s helped developed have applications that go beyond space exploration, such as factory floors here on Earth.
      With new missions on the horizon and smarter sensors in development, Hunter is still pushing the boundaries of what NASA technology can do. Whether it’s Titan, the surface of Venus, or somewhere we haven’t dreamed of yet, this work could help shape the way to get there. 

      Advanced materials research to make travel safer

      Advanced materials, such as foams and composites, are key to unlocking the next generation of manufacturing. From space exploration to industry, Erik Weiser spent years contributing his expertise to the development of polymers, ceramics, metals, nanomaterials, and more. He is named on more than 20 patents. During this time, he provided his foam expertise to the Space Shuttle Columbia accident investigation, the Shuttle Discovery Return-to-Flight Investigation and numerous teams geared toward improving the safety of the shuttle.  
      Today, Weiser serves as director of the Facilities and Real Estate Division at NASA Headquarters, overseeing the foundation of NASA’s missions. Whether it’s advancing research or optimizing real estate across the agency, he’s helping launch the future, one facility at a time.

      Want to learn more about NASA’s game changing innovations? Visit the NASA Inventors Hall of Fame.
      Read More Share
      Details
      Last Updated May 09, 2025 Related Terms
      Technology Technology Transfer Technology Transfer & Spinoffs Explore More
      3 min read Key Portion of NASA’s Roman Space Telescope Clears Thermal Vacuum Test
      Article 2 days ago 4 min read NASA Enables SPHEREx Data Return Through Commercial Partnership
      Article 3 days ago 6 min read NASA Data Helps Map Tiny Plankton That Feed Giant Right Whales
      In the waters off New England, one of Earth’s rarest mammals swims slowly, mouth agape.…
      Article 4 days ago Keep Exploring Discover More Topics From NASA
      Missions
      Humans in Space
      Climate Change
      Solar System
      View the full article
    • By European Space Agency
      Image: Part of the Italian island of Sardinia is featured in this image captured by the Copernicus Sentinel-2 mission. View the full article
    • By NASA
      5 min read
      Preparations for Next Moonwalk Simulations Underway (and Underwater)
      Editor’s Note: The following is one of three related articles about the NASA Data Acquisition System and related efforts. Please visit Stennis News – NASA to access accompanying articles.
      A blended team of NASA personnel and contractors support ongoing development and operation of the NASA Data Acquisition System at NASA’s Stennis Space Center. Team members include, left to right: Andrew Graves (NASA), Shane Cravens (Syncom Space Services), Peggi Marshall (Syncom Space Services), Nicholas Payton Karno (Syncom Space Services), Alex Elliot (NASA), Kris Mobbs (NASA), Brandon Carver (NASA), Richard Smith (Syncom Space Services), and David Carver (NASA)NASA/Danny Nowlin Members of the NASA Data Acquisition System team at NASA’s Stennis Space Center evaluate system hardware for use in monitoring and collecting propulsion test data at the site.NASA/Danny Nowlin NASA software engineer Alex Elliot, right, and Syncom Space Services software engineer Peggi Marshall fine-tune data acquisition equipment at NASA’s Stennis Space Center by adjusting an oscilloscope to capture precise measurements. NASA/Danny Nowlin Syncom Space Services software test engineer Nicholas Payton Karno monitors a lab console at NASA’s Stennis Space Center displaying video footage of an RS-25 engine gimbal test, alongside data acquisition screens showing lab measurements. NASA/Danny Nowlin Just as a steady heartbeat is critical to staying alive, propulsion test data is vital to ensure engines and systems perform flawlessly.
      The accuracy of the data produced during hot fire tests at NASA’s Stennis Space Center near Bay St. Louis, Mississippi, tells the performance story.
      So, when NASA needed a standardized way to collect hot fire data across test facilities, an onsite team created an adaptable software tool to do it.
      “The NASA Data Acquisition System (NDAS) developed at NASA Stennis is a forward-thinking solution,” said David Carver, acting chief of the Office of Test Data and Information Management. “It has unified NASA’s rocket propulsion testing under an adaptable software suite to meet needs with room for future expansion, both within NASA and potentially beyond.”
      Before NDAS, contractors conducting test projects used various proprietary tools to gather performance data, which made cross-collaboration difficult. NDAS takes a one-size-fits-all approach, providing NASA with its own system to ensure consistency.
      “Test teams in the past had to develop their own software tools, but now, they can focus on propulsion testing while the NDAS team focuses on developing the software that collects data,” said Carver.
      A more efficient workflow has followed since the software system is designed to work with any test hardware. It allows engineers to seamlessly work between test areas, even when upgrades have been made and hardware has changed, to support hot fire requirements for the agency and commercial customers.
      With the backing and resources of the NASA Rocket Propulsion Test (RPT) Program Office, a blended team of NASA personnel and contractors began developing NDAS in 2011 as part of the agency’s move to resume control of test operations at NASA Stennis. Commercial entities had conducted the operations on NASA’s behalf for several decades.
      The NASA Stennis team wrote the NDAS software code with modular components that function independently and can be updated to meet the needs of each test facility. The team used LabVIEW, a graphical platform that allows developers to build software visually rather than using traditional text-based code.
      Syncom Space Services software engineer Richard Smith, front, analyzes test results using the NASA Data Acquisition System Displays interface at NASA’s Stennis Space Center while NASA software engineer Brandon Carver actively tests and develops laboratory equipment. NASA/Danny Nowlin NASA engineers, from left to right, Tristan Mooney, Steven Helmstetter Chase Aubry, and Christoffer Barnett-Woods are shown in the E-1 Test Control Center where the NASA Data Acquisition System is utilized for propulsion test activities. NASA/Danny Nowlin NASA engineers Steven Helmstetter, Christoffer Barnett-Woods, and Tristan Mooney perform checkouts on a large data acquisition system for the E-1 Test Stand at NASA’s Stennis Space Center. The data acquisition hardware, which supports testing for E Test Complex commercial customers, is controlled by NASA Data Acquisition System software that allows engineers to view real-time data while troubleshooting hardware configuration.NASA/Danny Nowlin NASA engineers Steven Helmstetter, left, and Tristan Mooney work with the NASA Data Acquisition System in the E-1 Test Control Center, where the system is utilized for propulsion test activities.NASA/Danny Nowlin “These were very good decisions by the original team looking toward the future,” said Joe Lacher, a previous NASA project manager. “LabVIEW was a new language and is now taught in colleges and widely used in industry. Making the program modular made it adaptable.”
      During propulsion tests, the NDAS system captures both high-speed and low-speed sensor data. The raw sensor data is converted into units for both real-time monitoring and post-test analysis.
      During non-test operations, the system monitors the facility and test article systems to help ensure the general health and safety of the facility and personnel.
      “Having quality software for instrumentation and data recording systems is critical and, in recent years, has become increasingly important,” said Tristan Mooney, NASA instrumentation engineer. “Long ago, the systems used less software, or even none at all. Amplifiers were configured with physical knobs, and data was recorded on tape or paper charts. Today, we use computers to configure, display, and store data for nearly everything.”
      Developers demonstrated the new system on the A-2 Test Stand in 2014 for the J-2X engine test project.
      From there, the team rolled it out on the Fred Haise Test Stand (formerly A-1), where it has been used for RS-25 engine testing since 2015. A year later, teams used NDAS on the Thad Cochran Test Stand (formerly B-2) in 2016 to support SLS (Space Launch System) Green Run testing for future Artemis missions.
      One of the project goals for the system is to provide a common user experience to drive consistency across test complexes and centers.
      Kris Mobbs, current NASA project manager for NDAS, said the system “really shined” during the core stage testing. “We ran 24-hour shifts, so we had people from across the test complex working on Green Run,” Mobbs said. “When the different shifts came to work, there was not a big transition needed. Using the software for troubleshooting, getting access to views, and seeing the measurements were very common activities, so the various teams did not have a lot of build-up time to support that test.”
      Following success at the larger test stands, teams started using NDAS in the E Test Complex in 2017, first at the E-2 Test Stand, then on the E-1 and E-3 stands in 2020.
      Growth of the project was “a little overwhelming,” Lacher recalled. The team maintained the software on active stands supporting tests, while also continuing to develop the software for other areas and their many unique requirements.
      Each request for change had to be tracked, implemented into the code, tested in the lab, then deployed and validated on the test stands.
      “This confluence of requirements tested my knowledge of every stand and its uniqueness,” said Lacher. “I had to understand the need, the effort to meet it, and then had to make decisions as to the priorities the team would work on first.”
      Creation of the data system and its ongoing updates have transformed into opportunities for growth among the NASA Stennis teams working together.
      “From a mechanical test operations perspective, NDAS has been a pretty easy system to learn,” said Derek Zacher, NASA test operations engineer. “The developers are responsive to the team’s ideas for improvement, and our experience has consistently improved with the changes that enable us to view our data in new ways.”
      Originally designed to support the RPT office at NASA Stennis, the software is expanding beyond south Mississippi to other test centers, attracting interest from various NASA programs and projects, and garnering attention from government agencies that require reliable and scalable data acquisition. “It can be adopted nearly anywhere, such as aerospace and defense, research and development institutions and more places, where data acquisition systems are needed,” said Mobbs. “It is an ever-evolving solution.”
      Read More Share
      Details
      Last Updated May 08, 2025 EditorNASA Stennis CommunicationsContactC. Lacy Thompsoncalvin.l.thompson@nasa.gov / (228) 688-3333LocationStennis Space Center Related Terms
      Stennis Space Center View the full article
    • By USH
      In a groundbreaking development, advances in quantum data analysis have led to a discovery no scientist could have foreseen. NASA’s deep space monitoring system, upgraded with a quantum processor designed to filter cosmic noise and decode interstellar signals, produced something startling: an image.
      A conceptual interpretation of the Voyager 1 image.
      But this wasn’t an input, a simulation, or a product of algorithmic imagination. It wasn’t the result of random noise or a misfired pattern recognition process. The quantum system returned a coherent, structured, and symmetrical image, undeniably artificial. And the data it derived from? None other than Voyager 1. 
      Renowned physicist Michio Kaku addressed the anomaly in a recent interview: “We may be witnessing the first whisper of a new intelligence, something not man-made, not terrestrial, and certainly not random.” 
      The image, reconstructed via entangled qubit networks, depicted a figure: humanoid in silhouette, yet composed of geometric segments that defied any known biological or mechanical blueprint. It seemed deliberately crafted to challenge human comprehension, alien, yet eerily familiar enough to spark recognition.
      Not long ago, NASA pushed the boundaries of computation by launching an experimental quantum computer, capable of processing vast, multidimensional data streams. But after this revelation, NASA abruptly shut down the system following the unexpected and unsettling incident, in 2023, though some believe the research continued in secret. 
      Meanwhile, Voyager 1—the most distant human-made object in space, still traveling beyond our solar system after 45 years—has been transmitting strange, inexplicable data. According to NASA engineers, the spacecraft’s Attitude Articulation and Control System (AACS) began sending signals that “do not reflect what’s actually happening onboard.” 
      Instead of useful telemetry, Voyager 1 has been broadcasting a puzzling sequence: a repeating pattern of ones and zeros. Initially dismissed as a glitch, engineers traced the anomaly to the Flight Data Subsystem (FDS), pinpointing a malfunctioning chip. Yet, despite their efforts, the signal persisted, a digital enigma from 24 billion kilometers away. 
      Is this merely a failing system showing its age? Or is something, or someone, intentionally altering the data? 
      What if this “error” is a message? And if so, who’s sending it?
        View the full article
    • By Space Force
      The two-week event, held at Vandenberg Space Force Base, focuses on strengthening international partnerships, enhancing operational collaboration and promoting responsible behavior in space.

      View the full article
  • Check out these Videos

×
×
  • Create New...