Probabilistic ashfall forecasting with Latin Hypercube Sampling

This library provides the codes to prepare and analyse probabilistic ashfall forecasts using Latin Hypercube Sampling and Numerical Weather Prediction Model ensembles. Producing the probabilistic forecasts requires an ashfall forecasting model like HYSPLIT. Details on this work can be found here.

Installation

Install the conda environment

cd ashfall_forecasts 
conda env create -f environment.yml

Install package in new environment

conda activate lhs 
python setup.py develop

Setup Jupyter notebook kernel:

python -m ipykernel install --user --name lhs
pip install kernda
kernda -o -y /path/to/jupyter/kernels/lhs/kernel.json

To find the path of your kernel.json file you can run:

jupyter --paths