FORCE v. 3.7.9

Release: 25.10.2022

  • FORCE L1AS

    • The previously deprecated program force-level1-landsat was replaced by an overhauled version of landsatlinks. Downloading Landsat data through FORCE is now possible again! USGS/M2M credentials are required for this. A big shoutout to Stefan Ernst for enabling this!!!

  • FORCE L2PS

    • Added the option to buffer cirrus clouds. Cirrus clouds are quite hard to detect reliably (see CMIX paper). As such, buffering cirrus clouds often results in flagging out huge areas. To avoid that, we somewhen took the decision to not go for a cirrus buffer. This remains the standard behaviour.

      However, it is now possible to enable a buffer by using a buffer size > 0 for the new parameter CIRRUS_BUFFER. Note, that both the cirrus buffer won’t be represented in the “cloud buffer” QAI bit, but will be subsumed in the cirrus cloud bit (similar to snow buffer). This is because the bit structure would need to be disruptively changed to accomodate for a dsitinction.

      Thanks to Max Freudenberg for bringing this up.

  • FORCE WVDB

    • Increased the maximum number of characters to read the LAADS app key. Thanks to Fabian Thiel for bringing this up.

  • FORCE HLPS

    • When implementing custom GDAL options, a bug was unknowingly introduced into FORCE HLPS. To be flexible with regards to as many output drivers as possible, i.e. both drivers with create and copy capabilities, the following strategy was implemented: a memory-resident image was compiled with the MEM driver, which was then copied to disc with the requested driver. Apparently, this results in 0-values (although nodata value is different) being written into the image if the top-blocks are skipped, e.g. because of processing masks being used. This was fixed by only create-MEM/copying when the requested driver does not have create capabilities.

      As a note (and this was likely the case before, too), create-MEM/copying does not seem to work with block-output (like in HLPS). If full files are written, it works, however (like the OVV images in L2PS, where the same routine is used).

      Thanks to Stefan Ernst for bringin my attention to this.

    • force-higher-level will now print the number of bytes that were input and output at the end of the runtime. This might help to identify why processes might end quickly without producing expected data. Thanks to Max Freudenberg and Stefan Ernst for suggestions and discussions.

    • Marcel Schwieder added missing Interpolation parameters to the parameterfile skeleton, thanks!

  • FORCE AUX

    • Max Freudenberg increased the speed of force-mosaic when many files are present, thanks!

    • Stefan Ernst fixed empty tiles after running force-cube, thanks!