Operational Cloud Detection Algorithm for NASA’s Harmonized Landsat Sentinel-2

Operational Cloud Detection Algorithm for NASA’s Harmonized Landsat Sentinel-2 is sponsored by NASA. Developing algorithms for cloud detection in satellite imagery to enhance environmental monitoring.

Harmonized Landsat and Sentinel-2 | NASA Earthdata We are in the process of migrating all NASA Earth science data sites into Earthdata from now until end of 2026. Not all NASA Earth science data and resources will appear here until then. Thank you for your patience as we make this transition.

Read about the Web Unification Project Harmonized Landsat and Sentinel-2 The Harmonized Landsat and Sentinel-2 (HLS) project is an extension of research conducted at NASA's Goddard Space Flight Center that takes input data from the joint NASA/USGS Landsat 8 and Landsat 9 satellites and the European Space Agency's Sentinel-2A, Sentinel-2B, and Sentinel-2C satellites to generate a harmonized, analysis-ready surface reflectance data product with observations every 1.

6 days. Dr. Christopher S. R.

Neigh The HLS project is a major outcome of the Satellite Needs Working Group assessment in 2016.

In that assessment, federal agencies and end users identified a need for more frequent Landsat-like observations to track short-term changes in vegetation and other land components to support agricultural monitoring and land cover classification at moderate to high resolution in both the visible and thermal components of the electromagnetic spectrum.

Spectral similarities between the Landsat 8 Operational Land Imager ( OLI ), the Landsat 9 OLI-2, and the Sentinel-2 Multispectral Imager ( MSI ) present an opportunity to harmonize data from these sensors to generate higher-frequency imagery products for land surface monitoring and applications.

Previous versions of HLS data products produced by the HLS Science Team at Goddard had limited spatial coverage — only covering North America and other select global locations. The current version of the HLS algorithm is a cloud-based software stack that expands the spatial coverage to include all land masses globally, outside of Antarctica.

Two data products are generated as part of the HLS project: the L30 data product generated with Landsat 8 and Landsat 9 data, and the S30 product generated using Sentinel-2 data. These data are available through Earthdata Search as well as through NASA's Land Processes Distributed Active Archive Center ( LP DAAC ). Feedback or questions about HLS data products can be made in the Earthdata Forum for HLS .

Improving Land Monitoring Capabilities HLS data products greatly improve current publicly available remote sensing land monitoring capabilities, particularly in terms of the frequency of land surface observations through time.

The harmonization of HLS ensures that the Landsat 8 and Landsat 9 collection (30-meter spatial resolution with a 16-day repeat period) and the Sentinel-2A/B/C collection (10 to 20-meter spatial resolution with a five-day repeat period) can be used as if they were a single collection. Through HLS, land surface observations can be acquired at an unprecedented 30-meter spatial resolution every 1. 6 days.

The frequent revisit times of HLS allow for time series of land surface applications at the field/plot scale, which supports a wide range of applications, including agricultural health, insect infestations, and natural hazard impacts. Data from the Sentinel-2C mission (launched in 2024) is now active and incorporated into HLS data products, further improving re-visit times and land monitoring capabilities.

A suite of HLS vegetation index (VI) products was introduced in February 2025 and low-latency HLS products are anticipated to be completed by the end of 2025. The temporal frequency of HLS revisit times has improved with the launch of each satellite. This virtual constellation harmonization achieved a temporal revisit from 1.

6 days in 2022 with four satellites and 1. 4 days in 2025 with five satellites. Credit: HLS Science Group.

The HLS project is a collaborative effort between NASA and other external agencies, including the USGS and ESA. At NASA, the HLS algorithm is developed by the HLS Science Team and is ported and operated by the Interagency Implementation and Advanced Concepts Team ( IMPACT ) located at NASA's Marshall Space Flight Center in Huntsville, AL.

The data products are archived at and distributed by NASA's Land Processes Distributed Active Archive Center ( LP DAAC ), which is a partnership between NASA and the USGS. Coordination between these partners on the data life cycle ensures that HLS data are high-quality (research quality once the provisional tag is removed), freely available, and accessible to end users with two to three day latency.

External HLS partners (USGS and ESA) provide access to the input data at scale for global HLS data to be produced within the latency targets identified above. In addition, USGS provides the source code for atmospheric correction so the Level 1 input top of atmosphere reflectance can be converted to the output L30 and S30 surface reflectance data products.

HLS filenames (i.e., the local granule ID) follow a naming convention that gives useful information regarding the specific product. In this example of a tile product, the filename HLS. S30.

T60HTE. 2022103T222539. v2.

0. B01. tif indicates: HLS.

S30 – Product Short Name T60HTE– MGRS Tile ID (T+5-digits) 2022103– Julian Date of Acquisition (YYYYDDD) T222539– Time of Acquisition (HHMMSS) v2. 0 – Collection Version B01 – Variable such as Spectral Band, Angle Band, or Vegetation Index tif – Data Format (Cloud Optimized GeoTIFF) The HLS Long Name (i.e., Collection-Level) convention also provides useful information regarding the product.

In this example for an HLSS30 dataset, all products belonging to the HLS Sentinel-2 Multi-Spectral Instrument Surface Reflectance Daily Global 30 m V 2. 0 collection have the following characteristics: Sentinel-2 Multi-Spectral Instrument – Instrument/Sensor Surface Reflectance – Geophysical Parameter Daily – Temporal Resolution 30 m – Spatial Resolution V 2.

0 – Collection Version The HLS tiling system is identical to the one used for Sentinel-2. The tiles’ dimensions are 109. 8 km squares with an overlap of 4,900 m on each side.

The system is aligned with the Military Grid Reference System (MGRS) , and its naming convention is derived from the UTM (Universal Transverse Mercator) system. The UTM system divides the Earth’s surface into 60 vertical zones. Each UTM zone has a vertical width of 6° of longitude and horizontal width of 8° of latitude, as shown in the map below.

Each UTM zone is subdivided into MGRS 110 x 110 km zones. HLS uses a processing chain involving several separate radiometric and geometric adjustments, with a goal of eliminating differences in retrieved surface reflectance arising solely from differences in instrumentation.

Input data products from Landsat 8/9 (Collection 2 Level 1T top-of-atmosphere reflectance or top-of-atmosphere apparent temperature) and Sentinel-2 (L1C top-of-atmosphere reflectance) are ingested for HLS processing. A series of radiometric and geometric corrections are applied to convert data to surface reflectance, adjust for BRDF differences, and adjust for spectral bandpass differences.

Two types of products are then generated: “HLSS30” and “HLSL30. ” These products have been radiometrically harmonized to the maximum extent and then gridded to a common 30-meter UTM basis using the Sentinel-2 tile system. The S30 and L30 products are resampled as needed to a common 30-meter resolution UTM projection and tiled using the Sentinel-2 Military Grid Reference System (MGRS) UTM grid.

A detailed description of methods applied for processing and harmonizing Landsat 8/9 and Sentinel-2 data is described in the graphic below. LP DAAC distributes both L30 and S30 products. S30: MSI harmonized surface reflectance is resampled to 30m into the Sentinel-2 tiling system and adjusted to Landsat 8/9 spectral response function.

L30: OLI harmonized surface reflectance and Top-of-Atmosphere (TOA) brightness temperature is resampled to 30m into the Sentinel-2 tiling system. HLS products have two sources of metadata: the embedded COG metadata, and the external ECS (generated by the EOSDIS Core System) metadata. The COG metadata contains valuable information, including global attributes and dataset specific attributes pertaining to the granule.

The ECS . met file is the external metadata file in XML format, which is delivered to the user along with the HLS product. It provides a subset of the COG metadata.

Key features of HLS metadata attributes include georeferencing information and spectral band and QA band attributes: The ULX represents the X-coordinate of the Upper-left corner of the Upper-left pixel. The ULY represents the Y-coordinate of the Upper-left corner of the Upper-left pixel. The NCOLS represents the number of columns.

The NROWS represents the number of rows. The DATASTRIP_ID is the datastrip name in the SAFE (Sentinel Standard Archive Format for Europe) file for the HLSS30 product. The HORIZONTAL_CS_CODE is the projection code in EPSG format.

The HORIZONTAL_CS_NAME is the projection name HLS Data Prep Scripts at LP DAAC Publication: The Harmonized Landsat and Sentinel-2 surface reflectance data set Earthdata Webinar: Advancing Science Capabilities with Data Harmonization: NASA's Harmonized Landsat Sentinel-2 Product Earthdata Blog: HLS: Going Global, Going Cloud Landsat 8/Sentinel-2 Revisit Visualization Operational Land Imager ( OLI ) The Operational Land Imager aboard Landsats 8 and 9 measures nine bands in the visible, near infrared, and shortwave infrared portions of the spectrum.

Sentinel-2 Multispectral Imager ( Sentinel-2 MSI ) The Multispectral Imager aboard the European Space Agency Sentinel-2 series of satellites supports Copernicus Land Monitoring studies.