Landsat Collection 2: Satellite Imagery, Surface Reflectance, and GIS Workflows
Landsat Collection 2 is the USGS and NASA reprocessed Landsat archive, not a single raster layer. It contains Level-1 imagery from Landsat 1 through Landsat 9, global scene-based Level-2 surface reflectance and surface temperature products from Landsat 4 through Landsat 9, U.S. Analysis Ready Data, and selected Level-3 science products. The collection improves geometric accuracy, processing consistency, metadata, and cloud-based access. Its imagery supports land-cover mapping, vegetation monitoring, water analysis, urban change detection, agriculture, wildfire assessment, terrain interpretation, and long-term environmental studies. Landsat Collection 2 products are generally public domain, but analysts must still distinguish between product types, acquisition dates, processing tiers, quality masks, spatial resolutions, and geographic coverage.
Landsat Collection 2 is one of the foundational Earth-observation resources in the GIS ecosystem. It provides a long-running, globally distributed record of the physical surface of Earth through multispectral and thermal satellite measurements. The important distinction is that Collection 2 is an organized family of products rather than one uniform dataset: users may work with calibrated Level-1 data, atmospherically corrected Level-2 surface reflectance, surface temperature, U.S. Analysis Ready Data tiles, or higher-level science products. That structure affects how the data should be searched, downloaded, prepared, compared, and interpreted.
How EDK fits
For: Landsat Collection 2 is intended for GIS analysts, remote-sensing researchers, environmental scientists, land managers, agricultural users, public agencies, educators, and developers building Earth-observation workflows. It is especially useful when an analysis needs a consistent historical archive, multispectral measurements, thermal information, or repeat observations over a known location.
Landsat Collection 2 supplies time-indexed observations of physical places rather than a complete description of everything present on the ground. Its value in a broader world model is therefore complementary: Landsat can provide evidence about surface conditions and change, while geographic features such as roads, buildings, waterways, parcels, infrastructure, administrative areas, or service locations provide the contextual structure needed to interpret those observations. A location-data workflow can use Landsat imagery to ask questions such as where vegetation has declined near roads, which buildings or parcels fall within areas of persistent surface-water occurrence, how land cover differs around industrial sites, or where construction and disturbance appear between observation dates. The imagery itself does not establish an address, ownership record, asset identity, or legal boundary. It is an observation layer that becomes more useful when aligned with searchable geographic features and analyzed through spatial filtering, proximity, zonal statistics, change detection, and temporal comparison.
- •Combine Landsat Collection 2 surface reflectance or surface-temperature observations with roads, buildings, waterways, land-use polygons, or administrative boundaries to calculate regional summaries, compare neighborhoods, or detect environmental change around physical assets.
- •Use time-filtered imagery to examine vegetation, water, burn scars, exposed soil, or urban expansion near known geographic features, while retaining acquisition dates, cloud masks, sensor information, and processing-tier metadata as part of the result.
- •Use Collection 2 as an observational layer for geographic search and computation: identify places meeting spectral or thermal conditions, rank areas by change magnitude, and aggregate pixel-level measurements within selected regions.
Dataset profile
Formats
Cloud Optimized GeoTIFF imagery
Product metadata and ancillary files
Quality-assessment bands and masks
CSV bulk metadata
XML bulk metadata
Apache Parquet bulk metadata
Google Earth Engine image collections
Access methods
USGS EarthExplorer interactive search and download
USGS GloVis scene browsing and access
USGS and AWS cloud access for operational products
USGS bulk metadata services
Google Earth Engine catalog collections
NASA AρρEEARS for supported analysis-ready workflows
Quality and limits
Clouds, cloud shadows, haze, smoke, snow, and seasonal conditions can make observations unsuitable for a particular analysis.
Collection 2 includes multiple sensors and product levels with different band configurations, resolutions, processing rules, and geographic coverage.
Surface-temperature products are not guaranteed to be populated for every scene and require review of processing metadata and quality information.
Thirty-meter-scale observations are not a complete or authoritative inventory of individual buildings, roads, addresses, parcels, or infrastructure assets.
Time-series comparisons can be distorted by sensor differences, phenology, illumination, atmospheric conditions, missing observations, and inconsistent masking.
Important decisions should be validated with higher-resolution imagery, field observations, or authoritative local data when available.
Analysis workflows
Choose Level-1, Level-2, ARD, or Level-3 products according to the measurement and geography required.
Search by area, date range, sensor, product level, processing tier, and cloud-related metadata.
Download or stream the imagery together with metadata and quality-assessment bands.
Apply documented scale factors and mask clouds, shadows, snow, saturation, and other invalid pixels.
Calculate spectral indices or thermal measures using product-specific band definitions and quality rules.
Create seasonal composites or time series with consistent sensor, product level, spatial reference, and observation criteria.
Join rasters to buildings, roads, waterways, parcels, administrative areas, infrastructure, or other geographic features.
Calculate zonal statistics, proximity relationships, regional rankings, change surfaces, or threshold-based candidate areas.
Validate material findings against independent imagery, local datasets, or field information before treating them as physical-world conclusions.
Value inside EDK
What Landsat Collection 2 contains
Collection 2 is the second major reprocessing of the Landsat archive. USGS describes improvements in data processing, algorithm development, geometric accuracy, metadata, and data access. The collection contains Level-1 data from Landsats 1 through 9. Level-1 products provide calibrated and geometrically processed image data organized as scenes, with spectral bands, metadata, and quality information.
Collection 2 also contains Level-2 science products from Landsat 4 through Landsat 9. The main scene-based Level-2 products are surface reflectance and surface temperature. Surface reflectance is intended to reduce the influence of atmospheric conditions when comparing land-surface observations. Surface temperature provides thermal information where the necessary inputs and processing conditions are available. Some scenes may contain surface-reflectance data without a usable surface-temperature result, so users should inspect product metadata and processing-level fields rather than assuming every band is populated.
- •Level-1 calibrated imagery and metadata
- •Level-2 surface reflectance and surface temperature
- •U.S. Analysis Ready Data tiles
- •Selected Level-3 science products
- •Quality-assessment bands and ancillary metadata
Geography, time coverage, and update cadence
The scene-based Landsat archive is global. USGS lists worldwide Level-1 coverage for Landsats 1–9 and worldwide scene-based Level-2 surface reflectance and surface-temperature products for Landsats 4–9. The historical record begins in 1972 for the Landsat program, although the available product type depends on the sensor generation and processing level.
Landsat Collection 2 is an operational archive rather than a dataset with one fixed periodic release. Newly acquired Landsat 9 Level-1 products are generally available approximately 4–6 hours after acquisition, according to USGS. Level-2 products are generated after scenes enter the relevant processing tier and required auxiliary data are available; USGS describes typical Level-2 availability within about 24 hours after that point. Landsat 8 and Landsat 9 have nominal 16-day revisit intervals individually, while combined missions can provide more frequent observations in suitable locations. Actual usable observations are less frequent when clouds, shadows, snow, haze, or other quality problems obscure the target.
- •Global scene-based Level-1 coverage
- •Global scene-based Level-2 coverage for Landsat 4–9
- •U.S.-focused ARD coverage
- •Historical observations dating to the Landsat program's 1972 start
- •Ongoing operational additions rather than one universal release date
Formats and access methods
USGS distributes Collection 2 through multiple access paths. EarthExplorer supports map-based searches by location, date, and dataset and can provide individual or bulk downloads. GloVis provides access to Landsat Collection 2 Level-1 scene products. USGS also documents cloud access through Amazon Web Services for operational products and provides machine-readable catalog and metadata services.
Collection 2 imagery is distributed in Cloud Optimized GeoTIFF formats for many products, accompanied by metadata and quality files. Product-specific format-control documents should be consulted because band names, resolutions, metadata fields, masks, and file organization differ between sensors and product levels. USGS also maintains bulk metadata in CSV, XML, and, according to a March 2026 update, Apache Parquet formats. The bulk metadata files are updated daily and are useful for programmatic scene discovery across the archive.
- •USGS EarthExplorer for interactive search and download
- •GloVis for scene browsing and Level-1 access
- •AWS cloud access for operational Collection 2 products
- •Cloud Optimized GeoTIFF imagery with metadata and quality files
- •CSV, XML, and Apache Parquet bulk metadata
- •Google Earth Engine catalog collections
Suitable GIS and remote-sensing workflows
A defensible workflow begins by selecting the product that matches the question. Use Level-1 data when you need calibrated imagery and intend to perform your own processing. Use Level-2 surface reflectance for many land-surface comparisons, vegetation indices, land-cover studies, and change analyses. Use surface-temperature bands for thermal studies only after checking whether the scene contains a valid temperature product and reviewing the relevant quality information. Use U.S. ARD when tiled, analysis-oriented processing is more important than global scene consistency.
For a single-place analysis, search by a study boundary, acquisition date range, sensor, processing tier, and cloud-related metadata. Download the imagery and quality bands, reproject or resample only when necessary, mask clouds and cloud shadows, and preserve the original metadata. For multi-date analysis, keep sensor and product levels consistent where possible. Calculate indices such as NDVI or NDWI only after applying the appropriate scale factors and masks documented for the selected product.
- •Filter scenes by location, date, sensor, tier, and cloud information
- •Apply documented scale factors and QA masks before calculating indices
- •Use consistent product levels and seasonal windows for time series
- •Join imagery to geographic polygons for zonal statistics and regional comparison
- •Validate important findings against higher-resolution imagery or authoritative local data
License, strengths, and tradeoffs
USGS states that Landsat data are available at no cost and that federally created Landsat datasets are in the public domain. Users should provide acknowledgment or credit to the USGS as the data source. Third-party platforms, cloud services, processing environments, and derived commercial products may have their own operational terms even when the underlying Landsat data are public domain.
The largest strength of Collection 2 is continuity. It combines a long historical record with standardized reprocessing, global scene availability, multispectral bands, thermal information, quality layers, and several access methods. That makes it practical for repeatable geographic analysis across large regions and long time periods.
- •Strength: decades-long, globally distributed Earth-observation archive
- •Strength: public-domain data with no-cost USGS access
- •Strength: multispectral, thermal, quality, metadata, and analysis-ready products
- •Tradeoff: product families and sensors are not interchangeable without care
- •Tradeoff: clouds, shadows, haze, snow, smoke, and seasonal conditions limit usable observations
- •Tradeoff: moderate spatial resolution is unsuitable for many individual-asset or legal-boundary decisions
Questions teams ask
Is Landsat Collection 2 one dataset or several datasets?
It is a family of related products. The collection includes Level-1 imagery, Level-2 surface reflectance and surface temperature, U.S. Analysis Ready Data, and selected Level-3 science products. The correct product depends on the analysis objective, geography, sensor, and required preprocessing.
What is the difference between Landsat Collection 2 Level-1 and Level-2?
Level-1 products provide calibrated and geometrically processed scene data with associated metadata and quality information. Level-2 products add science-oriented processing such as surface reflectance and, where supported, surface temperature. Level-2 availability and populated bands depend on the sensor, scene, auxiliary inputs, and processing conditions.
Can Landsat Collection 2 be used for building or road detection?
It can support regional land-cover, disturbance, impervious-surface, and change analyses, but its resolution and observation conditions generally make it unsuitable as a definitive inventory of individual buildings or roads. Feature-level interpretation should be checked against higher-resolution imagery or authoritative local geographic data.
How often is Landsat Collection 2 updated?
There is no single update cadence for the entire collection. New operational scenes are added as satellites acquire and process them. USGS reports approximately 4–6 hours for newly acquired Landsat 9 Level-1 products, with Level-2 products following after tiering and required processing. Bulk metadata files are updated daily.
Keep exploring
landsat collection 2 analysis workflow
A workflow-focused companion for filtering scenes, applying QA masks, calculating indices, and comparing observations over time.
landsat collection 2 download and formats
A companion focused on EarthExplorer, cloud access, Cloud Optimized GeoTIFF products, metadata, and machine-readable formats.
global surface water analysis workflow
A related Earth-observation workflow for combining long-term raster observations with geographic features and regional statistics.
Sources
Landsat Collection 2
Primary USGS overview of Collection 2 contents, improvements, product generation timing, data policy, and documentation.
Open sourceLandsat Data Access
Primary USGS documentation for EarthExplorer, GloVis, AWS cloud access, ARD, Level-1, Level-2, and Level-3 product access.
Open sourceLandsat Data Processing, Archive, and Distribution
Primary USGS description of product generation, archive management, Level-1 and Level-2 availability, and distribution.
Open sourceLandsat Project Documents
Primary index for Collection 2 format-control books, science-product guides, ARD documentation, and product-specific technical references.
Open sourceLandsat Collections in Earth Engine
Official Google Earth Engine catalog overview describing Collection 2 availability by Landsat mission and the related image collections.
Open sourceLandsat Bulk Metadata Now Accessible in New Format
USGS announcement documenting daily bulk metadata updates and the addition of Apache Parquet alongside CSV and XML.
Open source