New HLS False Color Composite Layer Added to FIRMS

The impacts of large wildfires don't end when their flames are extinguished and the smoke clears. Beyond the loss of wildlife habitat, the combination of ash, charred soil, and loss of vegetation in burned areas can lead to increased erosion and runoff, and in extreme cases, flooding, mudslides, and the influx of sediments in reservoirs, which can imperil community water supplies.

Although ground-based sensors and aerial surveys may enable assessments of post-fire impacts in some locations, these methods may not be feasible in more remote areas, or in instances where a fire has burned hundreds of square miles. Using satellites to monitor burned areas is a more practical approach, but often the imagery from instruments like the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA's Aqua and Terra satellites and the Visible Infrared Imaging Radiometer Suite (VIIRS) instrument aboard the joint NASA-NOAA Suomi National Polar-orbiting Partnership and NOAA-20 satellites lacks the resolution sufficient for detailed evaluations of burned areas.

Now, with the recent addition of Harmonized Landsat Sentinel-2 (HLS) short-wave infrared false color composite imagery in the global and US/Canada instances of NASA's Fire Information for Resource Management System (FIRMS), that has changed.

"The HLS short-wave infrared false color composite imagery layer provides greater resolution and additional contrast for identifying and monitoring burned areas," said Dr. Brian Freitag, HLS Project Manager with the Interagency Implementation and Advanced Concepts Team (IMPACT) at NASA's Marshall Space Flight Center in Huntsville, Alabama. "If you were to look at just the standard true-color imagery layer over an active fire, a lot of what is happening on the ground is going to be obscured by a smoke plume. By changing the channels that we're putting into the red-green-blue composite, we're allowing users to see through the smoke to the surface and get a better glimpse of where the active fire front is."

The IMPACT team produced the global HLS dataset, which comprises the Landsat 30-meter (HLSL30) and Sentinel-2 30-meter (HLSS30) products, and works to ensure the quality of HLS data. It also produced the HLS false color composite imagery, which was developed at the request of the U.S. Forest Service.

HLS false color composite imagery is created with data from bands 7, 5, and 4 of the Operational Land Imager (OLI) and OLI-2 aboard the U.S. Geological Survey's Landsat 8 and 9 satellites, respectively, and bands 12, 8a, and 4 of the Multi-Spectral Instrument (MSI) aboard the ESA (European Space Agency) Sentinel 2A and 2B satellites. It is designed to help wildland fire personnel better delineate burned areas using short-wave and near infrared wavelengths undetectable by the human eye, and the result is high-resolution imagery with burned areas that appear almost brick red, making them easier to see on the landscape.

"The Forest Service uses false color composite imagery from MODIS, but it's one-kilometer [resolution], so when you have really large fires [the imagery] gets pixelated when you zoom in and try to see the active fire front or what the burn scar looks like," Freitag said. "HLS products bring it down to a 30-meter resolution, which gives users a higher-resolution look at the surface and the features of interest."

Brad Quayle, a remote sensing and GIS specialist with the Forest Service's Geospatial Technology and Applications Center (GTAC) agrees.

"From an active fire management standpoint, the HLS false color composite data are helpful in that they provide visualization of active fire fronts at a relatively higher spatial resolution at the time of the Landsat 8/9 or Sentinel 2A/2B overpass, which is helpful [for] verifying/validating the active fire detection data detected with coarser resolution sensors like MODIS," he said. "Also, the burn scars from prior years are discernible in these images, which can help inform fire managers where those previous fire areas may impede the spread of a current fire."

HLS false color composite data are also useful for monitoring the recovery of burned areas, Quayle said.

"The data also complement the false color composite MODIS and VIIRS imagery by providing higher fidelity imagery to view and discern fire impacts and damage on the landscape and, to a qualitative degree, how much those areas have recovered."

Further, the heightened contrast of surface features offered by false color composite imagery has applications beyond wildland fire.

"If you're looking at snowpack, it can be difficult to distinguish between cloud and snow in true color images, but the false color composite imagery layer shows snow or ice in blue and clouds appear in white," Freitag said. "It can also be used to detect flooding or bodies of water, which appear very dark, almost black."

The HLS project produces seamless, harmonized surface reflectance data from the OLI and MSI instruments aboard Landsat and Sentinel-2 Earth-observing satellites, providing a global observation of Earth's surface at a 30-meter spatial resolution every 2–3 days. In this case, however, "harmonized" does not mean "blended;" rather, it refers to the use of an algorithm that adjusts the data from each instrument so that the products made from them can be used interchangeably.

Although users can explore HLS true color Imagery in NASA Worldview and download HLS products from NASA's Land Processes Distributed Active Archive Center (LP DAAC) and NASA Earthdata Search, the false color composite image service is currently only available through FIRMS.

FIRMS, a product of NASA's Land, Atmosphere Near Real-time Capability for EOS (LANCE), distributes near real-time active fire data within 3 hours of a satellite observation, and its data have been integrated into NASA's Disasters Mapping Portal and used to generate maps of fire activity, position, perimeter, smoke direction, and other information critical to the fire response. FIRMS US/Canada, a more recent instance of FIRMS developed through an agreement between the U.S. Forest Service's GTAC and NASA, was created to modernize and optimize the Forest Service's distribution of active fire information by leveraging LANCE's web-based active fire mapping tools and capabilities for disseminating data, products, and services.

Like the original FIRMS, FIRMS US/Canada provides active fire imagery that can be viewed on an interactive fire map and current and historical corrected reflectance imagery from NASA and NOAA satellites. It also meets recently developed Forest Service requirements by offering additional contextual layers and enhancements, including the depiction of active fire fronts complete with time-since-detection information, incident locations, and ownership boundaries for current large fires in the U.S. and Canada, and National Weather Service Fire weather forecasts and red flag warning areas.

Users should note that unlike other satellite imagery layers in FIRMS, which are provided through NASA's Global Imagery Browse Services (GIBS), the HLS false color composite imagery is generated on demand through IMPACT, therefore it may take slightly longer to display than the HLS true color imagery through GIBS.

Users should also be aware of the latency associated with HLS products, Freitag said.

"To generate them, we take the Top-of-Atmosphere data from Landsat 8/9 and Sentinel-2 and we do an atmospheric correction to get surface reflectance. The data that allow us to do that atmospheric correction usually comes in at about two-to-three days latency," said Freitag. "So, in terms of active fire monitoring, or at least near real-time processing, HLS is not necessarily the best application, but it is useful for people on the ground to understand where the fire has come from and then monitor the recovery of the burned area."

Freitag acknowledges that, because full HLS global coverage is achieved roughly every two-to-three days, it's not always clear what observations will be available on a given day. To remedy that, users are encouraged to turn on the HLS true color imagery as well as the orbital paths for Landsat 8 and 9 and Sentinel 2A and 2B so they can see where the false color composite observations will be available. (Note: Data from Landsat 9 were incorporated into the HLS L30 product on May 17, 2022, in the forward mode. That means Landsat 9 data are available only from May 17 to the present. Freitag's colleagues are processing Landsat 9 data back to January 1, 2022, and plan to add it into the HLS data holdings in the next few months.)

"This should at least give users an idea of where data are expected on a given day; then they can zoom in and look at the data," he said.

And speaking of zooming in, the HLS false color composite data are not available at a global zoom level in FIRMS.

"HLS false color composite imagery supports zoom levels 10-to-14 in the FIRMS US/Canada interface, which means [users] have to get closer to the state level before they can actually start seeing the false color composite layer visualized," Freitag said. "A reason for that is if you're trying to pull 1,000 tiles, it's going to be very slow and that doesn't result in a good user experience. We've tried to find a balance where users will get visuals in a reasonable timeframe and be able to visualize an area large enough for fire monitoring."

Freitag's use of the term "tiles" refers to the HLS false color composite imagery tiles created dynamically in the cloud as clients request them. HLS is the first NASA Earth observation product fully processed and distributed in the cloud and it leverages several cloud-native standards. For example, products are distributed as Cloud Optimized GeoTIFFs (or COGs), which can be used in GIS applications, and include SpatioTemporal Asset Catalog (STAC) metadata. These open cloud-native standards offer exciting new approaches to delivering insights from the data through on-demand applications such as FIRMS. (Note: For a detailed description of how the tiles are created, see Getting a FIRMS Grip on Forest Fires on the Earthdata blog.)

"We've developed our entire production workflow and data distribution in the cloud," said Freitag. "We've generated the products as COGs, which allowed us to develop the dynamic tiler that has been integrated into FIRMS. The fact that the data are cloud-native, stored in the cloud, and served from the cloud, is significant, as it allows users to dynamically grab that data on the fly and generate these false color composites."

It's also significant in that, by being cloud-native, HLS is serving as something of a test case for NASA as it works to transition its Earth science data and data systems into the cloud and develop the next generation of tools and services for visualizing Earth observation data.

"We're expecting larger volumes of data over the next few years with NASA's joint mission with the Indian Space Research Organization (NISAR) and the Surface Water and Ocean Topography (SWOT) mission, both of which are going to be on the order of 20 to 50 TB of data per day. So, having this data physically in a location somewhere within NASA's existing data systems is likely to become a bit unwieldy," Freitag said. "We're starting to migrate all that processing and all of that data storage into the cloud and that's really what we're trying to do with HLS—to be that pioneering mission for future cloud-based data production and visualization services at NASA."