STAC

STAC#

STAC - Spatio-Temporal Asset Catalog

STAC - background#

  • Started in 2018, rapidly developing

  • New de facto metadata and search standard

  • Describes datasets at the level of individual files

  • It is most commonly used for remote sensing data, but it is suitable for any data with time and location information.

  • Users: ESA, USGS, Microsoft Planetary computer, Google Earth Engine

  • In Finland: FMI and CSC.

STAC concepts#

Search with STAC API#

  • Suitable also for very big datasets

  • Main focus on Item level

  • Search criteria:

    • Collection

    • Location: point, bbox, GeoJSON polygon

    • Time

    • Optional other Item values, for example cloud coverage

Tools for working with STAC#

  • In web browser: STAC Browser, STAC Index

  • QGIS: STAC plugin

  • Python: pystac-client, stackstac, xarray and dask

  • R: rstac, gdalcubes

  • PDAL: STAC reader

  • ArcGIS for Python API

  • Java, Julia, Ruby, Scala…

CSC Paituli STAC, Finnish spatial datasets#

  • ~100 different datasets, inlcuding:

  • Paituli raster datasets

    • LUKE, erosion risk maps

    • LUKE, topographic wetness index

    • LUKE, snow damage risk

    • NLS, orthophoto and infrared orthophoto*, 1996->

    • NLS, basic and topographic maps*, 2005->

    • FMI, historic weather in 10km grids*: min/mean/max temperature, precipitation, snow, sea level air pressure, humidity, radiotion, 1961->

  • GeoPortti Geocubes datasets

    • Finnish Forest center, Forest stand class

    • FMI, Wind

    • GTK, Superficial deposits

    • NLS, Orthophoto*

    • MAVI, Field parcels

    • SYKE, Corine Land Cover

    • SYKE, Vegetation height

  • FMI Tuulituhohaukka

    • ESA/FMI, Sentinel-1 daily and 11 days backscatter mosaics: VV and VH polarisation.

    • ESA/FMI, Sentinel-2 11-days and annual surface reflectance mosaics.

    • ESA/SYKE, Sentinel-2 monthly index mosaics: NDVI, NDBI, NDMI, NDSI, NDTI.

    • ESA, Sentinel-1 backscatter tiles: VV and VH polarisation.

    • USGS/SYKE, Landsat (4 and 5) yearly index mosaics: NDVI, NDBI, NDMI, NDSI, NDTI.

    • NLS, Digital terrain model products: DTM, aspect, slope.

    • Finnish Forest center, Canopy height model.

    • LUKE, Multi-source forest inventory products.

    • LUKE, Forest wind damage risk map.

    • FMI, Daily wind damage risk map.

  • ESA, Sentinel-2 products, processed to Level-2A (Surface Reflectance), a selection of mostly cloud-free products from Finland. Downloaded to CSC Allas by Maria Yli-Heikkilä (LUKE), Arttu Kivimäki (NLS/FGI) and Matias Heino (Aalto).

* These datasets have several bands in one file, Python stackstac does not support it, but search works.

Next steps