Tools developed as part of the SICE project at GEUS (Geological Survey of Denmark and Greenland).

Table of Contents

• Development Environment
• Tools description
  • extract_esalc.py
  • extract_arcticdem.py
  • get_IBOAR.py
  • SICE_products_availability.py
  • get_correlations.py
  • sice_tools_gui.py
  • SCDA.py
  • SCDA.xml
  • sicepy_multiprocessing.py
  • S3_wrapper.sh
  • S3_NRT.sh
  • SICE_processing.py

Development Environment

• These tools have been developed using a conda virtual environment that can be identically recreated. To this end, create a new environment using ./SICE_tools.yml as below:

  conda env create -f SICE_tools.yml
      

  ResolvePackageNotFound error can be raised. In that case, run conda env export --no-builds > SICE_tools.yml instead.
  Then, run conda activate SICE_tools to activate this new virtual environment.

• The repository can also be run interactively on Binder, without any download:

  

Tools description

extract_esalc.py

• Clips, reprojects and compresses ESA global Land Cover (ESALC) products (downloadable here) for a given region. Converts the resulting .tif file to be usable in the SICE toolchain.
• Iceland, Svalbard, FransJosefLand, NovayaZemlya, SevernayaZemlya, JanMayen, NorthernArcticCanada, SouthernArcticCanada, Norway, Beaufort and AntarcticPeninsula are currently implemented.
• Results have been merged with the master branch of the SICE toolchain and can be found here.
• A description of the 22 ESA global Land Cover (ESALC) products can be found here.

extract_arcticdem.py

• Clips ArcticDEM (downloadable here) derived slopes and slope aspects for a given region based on a mask.
• Slopes and slope aspects have been computed using the SNAP Slope Calculation operator.
• Masks have been generated using ./extract_esalc.py.

get_IBOAR.py

• Computes Effective Solar Zenith Angles (SZA) and Viewing Zenith Angles (VZA) based on s3_tools.
• The Intrinsic Bottom of Atmosphere Reflectance (IBOAR) is calculated for a given scene and given bands.
• Uses ArcticDEM derived slopes and slope aspects generated using ./extract_arcticdem.py and Rayleigh corrected Bottom of atmosphere Reflectances (BRR) using the SNAP Rayleigh Correction Processor.
• This code has been applied to Top of Atmosphere (TOA) reflectances in the SICE toolchain to implement a slope correction for the albedo and the snow grain diameter.

SICE_products_availability.py

• Checks the availability of the SICE toolchain products using a list or a csv file containing the product names.
• Option to run the functions using multiprocessing to drastically decrease computation time.
• Outputs can be visualised through a figure and/or saved in a csv file.

get_correlations.py

• Computes the correlation between two variables stored in rasters over a given specific area.
• The default variables are the Snow Grain Diameter (SGD) computed by the SICE toolchain and the percentage of variations between the Bottom Of Atmosphere Reflectance (BOAR) and the Intrinsic Bottom Of Atmosphere Reflectance (IBOAR) determined by ./get_IBOAR.py. This example can be easily modified with other variables and another function to apply.
• The function is run using multiprocessing to drastically decrease computation time.

sice_tools_gui.py

• Basis for a simple GUI to play interactively with different tools.
• A file open dialog, a matplotlib-hosted visualisation conserving projections (using rasterio), a 1D profile creation as well as a file save dialog are currently implemented.

SCDA.py

• Implementation of the Simple Cloud Detection Algorithm (SCDA) v2.0 using SLSTR bands, described in Fig. 5 of Metsämäki et al, 2015.
  METSÄMÄKI, Sari, PULLIAINEN, Jouni, SALMINEN, Miia, et al. Introduction to GlobSnow Snow Extent products with considerations for accuracy assessment. Remote Sensing of Environment, 2015, vol. 156, p. 96-108.
• v1.4 is also implemented based on the GlobSnow Technical note 2: Cloud Detection Algorithm SCDA.
• Channel S5 is calibrated as indicated in the Sentinel-3 Product Notice - SLSTR.
• The original syntax has been preserved to easily link back to sources.
• The algorithm can be run using multiprocessing to drastically decrease computation time.
• This code has been merged with the master branch of the SICE toolchain and can be found here.

SCDA.xml

• Extracts, resamples and reprojects SLSTR bands needed for ./SCDA.py.
• This file has been merged with S3.xml in the the master branch of the SICE toolchain.

sicepy_multiprocessing.py

• Runs sice.py, part of the SICE toolchain, using python multiprocessing with different strategies depending on whether the user runs one or multiple years and dates.

S3_wrapper.sh

• Wrapper of the SICE toolchain containing the modifications needed to run the option associated with ./sicepy_multiprocessing.py.

S3_NRT.sh

• Wrapper of the SICE toolchain modified to run Near Real-Time (NRT) over the 11 implemented arctic regions (Greenland, Iceland, Svalbard, Novaya Zemlya, Severnaya Zemlya, Frans Josef Land, Northern Arctic Canada, Southern Arctic Canada, Jan Mayen, Norway, Beaufort).
• Example of a Cron job:

  # m h  dom mon dow   command
  00 12 * * * /bin/bash -c "/path/to/wrapper/S3_NRT.sh" > /path/to/log/log_NRT.txt
      

SICE_processing.py

• Processes albedo outputs of the SICE toolchain (developed at v1.4).
• Computes an empirical Broandband Albedo (BBA) from OLCI Top of Atmosphere reflectances (r_TOA), further combined with planar shortwave broadband albedo when the latter is below bare ice albedo (0.565).
• Applies a temporal de-noising, smoothing and gap-filling modified after Box et al, 2017.
  Box, J. E., van As, D., & Steffen, K. (2017). Greenland, Canadian and Icelandic land-ice albedo grids (2000–2016). GEUS Bulletin, 38, 53-56.
• All steps except gap-filling are run using multiprocessing to drastically decrease computation time.