Published on 24 September 2024

swissEO S2-SR

Optical satellite data (Sentinel-2) showing land surface reflectance (SR) for all Sentinel-2 channels at spatial resolutions of 10 m, 20 m and 60 m. The product is currently available in two versions (see note on the transition phase below).

Note on the 2026 transition phase: The swissEO S2-SR product is currently available in two versions. The previous version, v100, will remain available until the end of 2026, after which it will be discontinued. All information and links relating to v100 can be found at the bottom of this page. The new version, v200, is available with immediate effect and includes expanded content as well as improved and faster processing methods. It is recommended that existing applications be migrated to v200 by the end of 2026.

Switzerland is imaged by four satellite orbits (see diagram below). The overlapping orbits provide a complete image of Switzerland approximately every three to five days; however, the usability of the data depends heavily on the weather conditions at the time of the flyover, as the imaging sensor cannot see through clouds. In addition to the data geocoding already applied, a co-registration process optimised for Switzerland is used to achieve sub-pixel positional accuracy, and the data is reprojected into a uniform coordinate system. Optimised quality layers are also provided, including masks for clouds (and cloud shadows) as well as topographical shadows.

Coverage of Swiss territory depending on the orbit of the satellite.

Data visualisation

The latest dataset (v200) can be viewed as a true-colour image in the BGDI map viewer via https://map.geo.admin.ch/?layers=ch.swisstopo.swisseo_s2-sr_v200. In the BGDI’s WMS, the dataset can be displayed as a true-colour image by using the layer ‘ch.swisstopo.swisseo_s2-sr_v200’.

Processing of a newly acquired dataset following its acquisition currently takes around 10–24 hours. The data acquisitions planned for the current and following week can be viewed via the calendar below, which is based on the acquisition schedules of the European Space Agency’s (ESA) Sentinel-2 satellites.

To view a time series, there is a catalogue or a machine-readable list of the acquired and processed datasets, as well as a further list of the days on which no data was acquired or cloud cover was too high (orbits 8 and 22: > 85%, orbits 108 and 65: > 95%), in which case the data is not processed. To display older datasets in the BGDI map viewer, a URL parameterisation must be carried out. In the link below, must be replaced with the timestamp of the data acquisition in the format ‘yyyy-mm-ddtHHMMSS’ (example: 2025-01-01t235959).

Via WMS

https://map.geo.admin.ch/index.html?layers=WMS||s2-sr_v200_<id>||https://wms.geo.admin.ch/?item=<id>||ch.swisstopo.swisseo_s2-sr_v200||1.3.0

For example, for the dataset dated 10 May 2024 at 10:20:21:

https://map.geo.admin.ch/index.html?layers=WMS||s2-sr_v100_2024-05-10t102021||https://wms.geo.admin.ch/?item=2024-05-10t102021||ch.swisstopo.swisseo_s2-sr_v200||1.3.0

Data availability

Download raw data

Raw data can be downloaded via the collection «ch.swisstopo.swisseo_s2-sr_v200» in the STAC interface of the BGDI:

• Download interface: swissEO S2-SR: Optical satellite data (Sentinel-2) (admin.ch)
• Download API documentation: REST Interface: Spatial Temporal Asset Catalog API (admin.ch)
• Accessing STAC via Python - Introduction (Jupyter Notebook)
• Image processing via Python using STAC (Jupyter Notebook)
• Streaming of swissEO products in QGIS

Specific examples:

• List of all available raw data as JSON: https://data.geo.admin.ch/api/stac/v0.9/collections/ch.swisstopo.swisseo_s2-sr_v200/items
• List of all available raw data within the quadrangle with coordinates 7.43, 46.95, 7.69, 47.10 as JSON: https://data.geo.admin.ch/api/stac/v0.9/collections/ch.swisstopo.swisseo_s2-sr_v200/items?bbox=7.43,46.95,7.69,47.10
• List of all available raw data within the quadrangle with the coordinates 7.43,46.95,7.69,47.10 and the time period: 1 April 2024 – 10 April 2024 https://data.geo.admin.ch/api/stac/v0.9/collections/ch.swisstopo.swisseo_s2-sr_v200/items?bbox=7.43.46.95.7.69.47.10&datetime=2024-04-01/2024-04-10

Additional information

• Technical details

  • Coordinate system: CH1903+ LV95 (EPSG:2056)
  • Resolution: 10m, 20m, 60m
  • Temporal resolution: For each flyover (approx. 2–3 times per week)
  • Availability: Orbits that are at least 85 per cent (orbit 8.22) or 95 per cent (orbit 108.65) cloud-free
  • Spatial coverage: Switzerland including adjacent lakes, the Principality of Liechtenstein, and a 5 km buffer zone
  • Data format: Cloud Optimised GeoTIFF (COG)
  • Interfaces: STAC, WMS
  • Scaling of reflectance values: scaling factor 0.0001, offset -0.1 (physical value = pixel value × 0.0001 - 0.1)
  • Source code: https://github.com/swisstopo/topo-satromo-v2

• Product contents

  The swissEO S2-SR v200 product contains 18 elements, all of which can be downloaded directly:*

  • AOT-60M: Aerosol Optical Thickness (AOT band) – 60 m. (Sample data)
  • B01-60M: Coastal Aerosol (Band 1) – 60 m. (Sample data)
  • B02-10M: Blue (Band 2) – 10 m. (Sample data)
  • B03-10M: Green (Band 3) – 10 m. (Sample data)
  • B04-10M: Red (Band 4) – 10 m. ( Sample data )
  • B05-20M: Red Edge 1 (Ribbon 5) – 20 m. ( Sample data )
  • B06-20M: Red Edge 2 (Roll 6) – 20 m. ( Pattern data )
  • B07-20M: Red Edge 3 (Volume 7) – 20 m. ( Pattern details )
  • B08-10M: NIR 1 (Band 8) – 10 m (sample data)
  • B8A-20M: NIR 2 (Band 8A) – 20 m. ( Sample data )
  • B09-60M: NIR 3 (Band 9) – 60 m. (Sample data)
  • B11-20M: SWIR 1 (Band 11) – 20 m. (Sample data)
  • B12-20M: SWIR 2 (Band 12) – 20 m. (sample data)
  • SCL-20M: Scene Classification Map – 20 m. Contains the Sentinel-2 scene classification (classes 0–11). (Sample data)
  • CLOUDMASK-10M: Cloud mask according to OmniCloudMask – 10 m. Contains cloud and cloud shadow classification (classes 0–3). (Sample data)
  • TERRAINMASK-10M: Solar incidence angle and shadow casting – 10 m. Contains the solar incidence angle of light (0–180°), calculated from the azimuth and zenith angle of the sun, as well as a shadow mask (value 200) for all shaded pixels based on swissSURFACE3D at the time of acquisition. Calculated using HORAYZON. (Sample data)
  • TCI-10M: True Colour Image – 10 m. Visually optimised true-colour image as a Cloud Optimised GeoTIFF, ready for direct use in mapping and web applications. (Sample data)
  • METADATA: JSON file containing all necessary information for each element of the product. (Sample data)
  • THUMBNAIL: A reduced-size PNG image of the satellite image for faster previewing. (Preview)

  * All sample data and previews refer to the dataset from 18 January 2020 at 10:43:51 as an example.

  Note on the scaling of reflectance values (v200): The pixel values of the spectral bands (B01–B12, AOT) are stored in a scaled format. The physical reflectance value is calculated as follows: Physical value = Pixel value × 0.0001 − 0.1. This step must be explicitly taken into account during data processing.

  Additional information on the content v200 (English)

  • swissEO S2-SR - Additional Content Info - v200

    PDF71.39 kB12 August 2026

• Detailed product information

  The data is based on the European Space Agency’s (ESA) Sentinel-2 L2A data. Sentinel-2 comprises three satellites (Sentinel-2A, Sentinel-2B and Sentinel-2C). The Sentinel-2 sensors acquire data across 13 spectral bands, ranging from visible light through the near-infrared to the short-wave infrared, with spatial resolutions of 10 m, 20 m and 60 m. For the ‘swissEO S2-SR’ v200 product, all available bands from the Copernicus Data Space are processed: B01 (Coastal Aerosol), B02, B03 and B04 (blue, green and red), B05, B06, B07 (red-edge bands), B08, B8A and B09 (near-infrared), B11, B12 (short-wave infrared), and AOT (Aerosol Optical Thickness). The product covers Switzerland and the Principality of Liechtenstein, including the adjacent lakes and a 5 km buffer zone around the national territories.

  The OmniCloudMask algorithm is used to identify clouds and cloud shadows. This neural network-based approach classifies clouds and cloud shadows directly using the Sentinel-2 reflectance bands, without requiring external reference data. The result is a classified cloud and cloud shadow mask (CLOUDMASK-10M) that identifies both direct cloud pixels (distinguishing between thick and thin clouds) and cloud shadows.

  The terrain shadow calculation provides the solar incidence angle for each pixel, as well as a precise shadow mask. The TERRAINMASK-10M band contains the following values:

  • 0–180: Solar angle of incidence in degrees (illuminated pixels). 0° corresponds to vertical illumination from above; values close to 90° correspond to extreme oblique illumination.
  • 200: Shadow (cast shadow). The shadow calculation is based on the swissSURFACE3D digital surface model and the position of the sun at the time of acquisition (deviation ±2.5 min). The calculation is performed using the HORAYZON library, which enables a pixel-accurate simulation of shadow casting.

  Images from the same orbital track (satellite orbit) over Switzerland are stitched together to form a mosaic. This optimises data consistency and reduces artefacts at the image edges. Where there are two orbital tracks from different satellites, two mosaics are provided, one for each pass.

  Co-registration is carried out by aligning the Sentinel-2 images with a SWISSIMAGE reference image resampled to 5 m. The AROSICS image alignment algorithm is used for this, which is based on the red spectral band (B4), as this band is particularly high in contrast and clearly depicts distinctive landscape features. The algorithm searches for matching image features between the current image and the reference image and uses these to calculate displacement vectors in the X and Y directions. The displacement vectors are applied to all supplied bands. The displacement vectors can be provided on request.

  The True Colour Image (TCI-10M) in v200 is a visually optimised true-colour image that has been specially processed to provide an appealing representation of the Swiss landscape. It is available as a Cloud Optimised GeoTIFF and can be integrated directly into map applications, web applications or used as a background image.

---

Note on the 2026 transition phase: The swissEO S2-SR product is currently available in two versions. The previous version, v100, will remain available until the end of 2026, after which it will be discontinued.

• Information – swissEO S2-SR v100 (until the end of 2026)

  Data visualisation and availability – v100 (until the end of 2026)

  The most recent dataset can be displayed as a true color image in the BGDI map viewer at https://map.geo.admin.ch/?layers=ch.swisstopo.swisseo_s2-sr_v100. In the BGDI WMS, this dataset can be displayed using the «ch.swisstopo.swisseo_s2-sr_v100» layer.

  It currently takes around three days to process a newly recorded data set after it has been recorded. To find out which data acquisitions are planned for the current and following week, please visit the website of the European Space Agency (ESA) to check the recording schedules of the Sentinel-2 satellites.

  In order to view a time series, there is a catalogue or a machine-readable list of the recorded and processed data sets, as well as a further list with the days on which no data was recorded or the cloud cover was greater than 80% and the data was therefore not processed. To display older data sets as TrueColor RGB images in the BGDI map viewer, a URL parameterisation must be carried out. The in the following link is replaced by the timestamp of the data acquisition in the format “yyyy-mm-ddtHHMMSS” (for example: 2025-01-01t235959).

  https://map.geo.admin.ch/index.html?layers=WMS||s2-sr_v100_||https://wms.geo.admin.ch/?item=||ch.swisstopo.swisseo_s2-sr_v100||1.3.0

  For example for the 2024-05-10t102021 dataset:

  https://map.geo.admin.ch/index.html?layers=WMS||s2-sr_v100_2024-05-10t102021||https://wms.geo.admin.ch/?item=2024-05-10t102021||ch.swisstopo.swisseo_s2-sr_v100||1.3.0

  Download raw data – v100 (until the end of 2026)

  Raw data can be downloaded via the collection «ch.swisstopo.swisseo_s2-sr_v100» in the STAC interface of the BGDI:

  • Download interface: swissEO S2-SR: Optical satellite data (Sentinel-2) (admin.ch)
  • Download API documentation: REST Interface: Spatial Temporal Asset Catalog API (admin.ch)
  • Google Earth Engine: S2-SR Image Collection [projects/satromo-prod/assets/col/S2_SR_HARMONIZED_SWISS]
  • Accessing STAC via Python - Introduction (Jupyter Notebook)
  • Image processing via Python using STAC (Jupyter Notebook)
  • Streaming of swissEO products in QGIS

  Technical details – v100 (until the end of 2026)

  • Coordinate system: CH1903+ LV95 (EPSG:2056)
  • Resolution: 10m
  • Temporal resolution: for each overflight (approx. 2-3 times per week)
  • Availability: Orbits that are at least 20% cloud-free.
  • Data Format: Cloud Optimized GeoTIFF (COG)
  • Interfaces: STAC, WMS
  • Source code used for the production: https://github.com/swisstopo/topo-satromo/

  Product contents – v100 (until the end of 2026)

  The swissEO S2-SR product contains seven elements all available for direct download:*

  • BANDS-10M: True color orthophoto of the RGBN Copenicus bands (B4, B3, B2, B8) with a 10m resolution. (Preview, Sample)
  • BANDS-20M: Red-Edge, Near Infrared and Shortwave Infrared of Copernicus bands (B5, B8A, B11) with a resolution of 20m. (Sample)
  • CLOUDPROBABILITY-10M: This raster contains the probability of cloud cover for each pixel in the satellite image. (Sample)
  • MASKS-10M: This raster contains masks for various features such as terrain shadows, clouds, and cloud shadows of the satellite image. (Sample)
  • METADATA: JSON file with all necessary information for every element of the product. (Sample)
  • REGISTRATION-10M: This raster contains registration information for the satellite data, including displacement in the x and y directions. (Sample)
  • THUMBNAIL: A resized JPEG image of the satellite image to enable a quicker preview. (Preview)

  * All samples and previews refer to the latest product update.

  Additional content information – v100 (until the end of 2026)

  • swissEO S2-SR - Additional content information

    PDF380.18 kB25 March 2025

  Detailed product information – v100 (until the end of 2026)

  The data basis is provided by Sentinel-2 L2A data from the European Space Agency (ESA). Sentinel-2 comprises three satellites (Sentinel-2A, Sentinel-2B and Sentinel-2C). The Sentinel-2 sensors capture data in 13 spectral bands of visible light across the near-infrared range to short-wave infrared with spatial resolutions of 10 m, 20 m and 60 m. For the ‘swissEO S2-SR’ product, bands B2, B3, B4 and B8 (blue, green, red and near infrared) are processed at a spatial resolution of 10 m, and bands B8A, B11 and B5 (narrow near infrared, shortwave infrared and red edge) at a spatial resolution of 20 m.

  The CloudScore+ data, specially developed for Sentinel-2 data, is used to identify clouds and cloud shadows. This algorithm calculates a spectral distance value for each pixel relative to a theoretically clear reference, enabling an assessment of cloud probability. The masking process identifies both direct cloud pixels and their shadow areas. Clouds above a threshold value of 40% cloud probability are masked. A lower threshold value of 20% is applied for cloud shadows, but only in the projection corridor of clouds that have already been identified. In order to detect semi-transparent cloud edges and eliminate isolated misclassifications, morphological filter operations are applied to the combined cloud and cloud shadow mask. Both the binary cloud and cloud shadow mask and the continuous cloud probability layer itself are provided, allowing you to define your own threshold values if required.

  Terrain shadow detection is performed using a two-step approach that takes into account both the illumination angle and direct shadow casting. In the first step, the illumination angle is calculated for each pixel, which is determined from the sun's zenith angle, the sun's azimuth, and the local slope and orientation. This angle indicates how directly the sun's rays hit the ground surface. 0° means perfect illumination from above, while high values towards 90° mean extreme oblique illumination. In a second step, direct shadows are identified using the swissSURFACE^3D digital surface model. The shadows are simulated depending on the position of the sun at the time of recording (deviation ±2.5 min). The identified direct shadows are marked with a value of 100 in the terrain shadow mask.

  Images from the same orbital track of a satellite over Switzerland are combined into a mosaic. This optimises data consistency and reduces artefacts at the edges of the images. If two orbital tracks from different satellites are available, two mosaics are provided, one for each orbital track.

  Co-registration is performed by comparing the Sentinel-2 images with a reference image (S2 Global Reference Image, GRI). An image matching algorithm based on the red spectral band (B4) is used, as this band is particularly high in contrast and clearly shows distinctive landscape features. The algorithm searches for matching image features between the current image and the reference image and calculates displacement vectors in the X and Y directions. The displacement vectors are applied to all delivered bands and are also made available for traceability purposes.

Terms of use

swissEO may be used, distributed and made accessible. Furthermore,it may be enriched and processed and also used commercially. A reference to the source is mandatory. In the case of digital or analogue representations and publications, as well as in the case of dissemination, the following source references must be attached in any case:

© swisstopo, Contains modified Copernicus Sentinel data [Year]

Further information on the terms of use can be found on the terms of use webpage.