Sources/Usage: Public Domain. View Media Details

The 22nd William T. Pecora Memorial Remote Sensing Symposium (Pecora 22) kicks-off today, October 24, 2022, in Denver, Colorado. Pecora 22 celebrates the 50th anniversary of the launch of the first Landsat satellite and the accomplishments that followed. The symposium focuses on all aspects of Earth observation, spanning scientific discoveries to operational applications and from sensors to decisions. Embracing innovations, discoveries, and 50 years of history, the Pecora 22 conference theme is Opening the Aperture to Innovation: Expanding Our Collective Understanding of a Changing Earth.  

USGS Earth Resources Observation and Science (EROS) Center staff will be sharing presentations and posters throughout the conference. Here are a few examples. 

Watching Land Change Nationally 

Every two to three years, USGS EROS releases a definitive National Land Cover Database (NLCD) based on Landsat imagery. On a pixel-by-pixel level, the data layers detail transformations of the United States’ land surface. “Satellite imagery is the beginning of everything. It provides the information,” said Suming Jin, the USGS NLCD Chief Scientist. Each database must be consistent and accurate, but also processed efficiently. 

The recent release of NLCD 2019 was created using composite images, which were based on selecting a clear observation with the minimal Euclidean distance to the virtual Median Value Point (MVP) that has the median value of all clear observations of a pixel across the compositing time period. This reduced the shadows and clouds found in the base imagery, which brings out better information. It worked well compared to the other compositing methods and reduced a lot of effort to process individual Landsat path/row images, Jin said, but composite imagery can still have some cloud, cloud shadow, and missing values. 

Currently, Jin is testing strategies for the next data release, NLCD 2021. Most recently, she’s exploring synthetic source imagery, which can be produced during the USGS Land Change Monitoring, Assessment, and Projection (LCMAP) initiative's Continuous Change Detection and Classification (CCDC) operational process and can be generated for any given date. By comparison to composite imagery, synthetic imagery can generally be produced more consistent with every pixel remaining cloud-free.  

Sources/Usage: Public Domain. View Media Details

Synthetic imagery does have its limitations. Jin compared the two image sources with various tests, including long-term and short-term change detection and different land cover types change detection. An example of composite and synthetic imagery can be seen here featuring White Oak City, Alabama. Synthetic imagery had more omission errors when compared to composite imagery, like not noticing subtle, gradual changes in a time series of a forested area. On the other hand, composite imagery had more commission errors, detecting change when there was none. Often synthetic imagery is poor at detecting snow/ice, water, and agriculture changes. 

Jin determined the best strategy was both imagery types. NLCD 2021 will use synthetic source imagery and composite imagery for change detection and land use classification. By using both, the next landcover database will be more accurate. 

Upgrades to Landsat Science and Data Products 

The USGS develops Landsat science products to reduce the amount of time users spend on processing data, shared Michelle Bouchard, a contractor and communications specialist for the USGS EROS Center Landsat Program Office. Her poster showcases “Long-term USGS Landsat Collection 2 Science Products,” including Level-2, Level-3, and U.S. Analysis Ready Data. These science products ensure continuity to the Landsat record, allowing users to access large datasets with a consistent, quality product. 

First made available in 2020, Collection 2 harnesses advancements in data processing, algorithm development, data distribution, and data access capabilities to enhance Landsat data products. The collection includes Landsat Level-1 data for all sensors since 1972 as well as global Level-2 surface reflectance and surface temperature scene-based products from 1982 to present.  

“Users should move to Collection 2 as soon as possible,” Bouchard shares. Collection 1 has not been updated with Landsat products since December 31, 2021 and is scheduled for removal at the end of 2022. 

Measuring Water Quality from Space 

Benjamin Page, a contractor and remote sensing scientist with USGS EROS Center, prepared an introduction to the new Provisional Landsat 8/9 Aquatic Reflectance (AR) Science Product. This AR science product allows users to measure water quality from space, said Page, which “help users evaluate changes in surface water environments.” These satellite observations can be used in conjunction with in-situ samples and running tests in the lab.  

Fifty years ago, Pecora and his contemporaries designed Landsat to use 30-meter spectral resolution to image the surface, not for aquatic environments. When Landsat’s sensors image water, the colors are dark with respect to the surrounding land pixels, not ideal for aquatic reflectance studies.  

Sources/Usage: Public Domain. View Media Details

However, Landsat and the decades of data it provides could be leveraged with the right algorithm, as available in the USGS Landsat AR science product. Researchers like Page refine AR algorithms to read these dark colors that Landsat 8 and Landsat 9 capture with the Operational Land Imager (OLI) sensor. Page shared that Landsat can “target inland waterbodies” like lakes and rivers that ocean-based satellites are unable to image. 

“The ultimate goal of the AR science product is to allow users to measure water quality from space,” said Page. “We are still in the standardization stages and need global verification.” 

Water monitoring communities can apply the Landsat Aquatic Reflectance product to evaluate water quality and water clarity through OLI-based aquatic reflectance. Understanding changes in aquatic reflectance across different wavelengths can infer locations of algae blooms, sediment discharge, and other public health risks.  

PECORA Landsat Presentations 

Additional presentations and posters about Landsat from USGS staff will be available. The following list shows a selection, times are in Mountain Daylight Time (MDT). For official dates and times, please reference the Pecora 22 schedule here. 

Technical Sessions on Tuesday, October 25 

• Chris Crawford, “Perspectives on USGS Landsat Processing and Future Needs,” 10:30 a.m. 

• Benjamin Page, “An Introduction to the Landsat 8/9 Collection 2 Provisional Aquatic Reflectance Science Product,” 10:45 a.m.   

• Suming Jin, “NLCD Change Detection Using CCDC Synthetic and Composite Imagery,” 10:45 a.m. 

• Matthew Rigge, “Rangeland Condition Monitoring Assessment and Projection (RCMAP): tracking ground cover patterns over a 1985-2021 time series,” 11:15 a.m.  

• Shannon Franks “Accuracy Analysis of Copernicus DEM and comparison to the current USGS DEM used in Landsat processing,” 11:15 a.m. 

• Cody Anderson, “Quantitative Metrics for Interoperability” for Analysis Ready Data (ARD),” 11:15 a.m. 

• Cody Anderson, “Landsat 9 Calibration Overview—Commissioning and Move to Operations,” 2:00 p.m. 

• Mike Choate, “An Overview on Landsat 9 Operational Geometric Characterization and Calibration Processes and Results,” 2:30 p.m.  

• Rajagopalan Rengarajan, “Comparing the Geometric Performance of Landsat 8 and Landsat 9 Satellites and Their Data Products,” 2:45 p.m.  

• Kate Fickas, “Sending Mixed Signals: Predicting Landsat Reflectance Patterns with Climate Data,” 1:30 p.m.  

• Terry Sohl, “Historical Landscape Reconstruction and Scenario-Based Modeling in LCMAP,” 2:00 p.m.  

• Emily Maddox, “Landsat 8/9 Under-Fly Surface Reflectance Validation,” 3:45 p.m.  

Poster Session on Tuesday, October 25 

• Michelle Bouchard, “Long-term USGS Landsat Science Products.” 

• Cole Krehbiel, “Land Change Monitoring, Assessment, and Projection (LCMAP): E-Learning Materials to Connect Users to Over Three Decades of Landsat-derived Land Cover Science Products.” 

Technical Sessions on Wednesday, October 26 

• Gabriel Senay, “Improving Evapotranspiration Estimation and Mapping for Global Field-Scale Applications,” 10:30 a.m.–12:00p.m.  

• Saeed Arab, “Modeling Snowmelt Runoff Using Landsat Fractional Snow-Covered Area,” 10:45 a.m.  

• Jacob Savoy, “Landsat Data Access Updates,” 3:30 p.m.  

• Holly Wilson, “Introduction to Using Landsat Data in the Cloud,” 3:45 p.m. 

For more information see the conference website at http://pecora22.org. Questions? Complete the form under "Contact" on the website or email the Pecora 22 Technical Program Committee at pecora@usgs.gov. 

Return to all Landsat Headlines