About

Snowdrift is a forecast model component for snowdrift prediction. It implements a novel snow drift forecast algorithm developed and documented by Vegsýn ehf. (Dr Skuli Thordarson) in 2013 and 2014.

The model was first implemented for routine numerical weather prediction at the Icelandic Meteorological Office in 2014, and has been running on the Icelandic LAM forecast ever since.

The original work behind the model was financed with grant support from the European Northern Perimetry Program (SNAPS project) and the Icelandic Roads and Coastal Administration. This new software implementation of the model has been commissioned by the Icelandic Meteorological Office in order to make the model more portable and configurable for different model inputs.

The project has been further supported by grant in 2021 from the Icelandic Roads and Coastal Administration as part of developing a novel road-weather forecast interface for road authorities.

The snowdrift algorithm is documented in the following reports:

Quickstart

install deps, if not already installed:

$ sudo apt-get install libeccodes-dev git feh
$ sudo pip3 install numpy bcolz pygrib pillow trollimage nose more_itertools

get the project and run tests:

$ git clone "https://gitlab.com/vedurstofan/snowdrift"
$ cd snowdrift
$ export PYTHONPATH=$PWD
# tests data will be downloaded to './tests/testdata/'
$ python3 -m "nose" -v ./tests/

generate some snowdrift forecast frames:

$ bin/snowdrift tests/testdata/harmonie* --out imgs/test --format PNG
# loop through the resulting images,
$ feh -D 0.3 imgs/test*
# generate a snowdrift forecast in GRIB format
$ bin/snowdrift tests/testdata/harmonie* --out snowdrift.grb --format GRIB

Install the snowdrift module on your platform:

$ sudo pip3 install .
# OR
$ sudo pip3 install "git+https://gitlab.com/vedurstofan/snowdrift"
# The CLI app should now be available on the command line
$ snowdrift -h