This is a python package that implements an ecosystem model of grassland carbon (C) biogeochemistry (DALEC-Grass). DALEC-Grass is integrated into a model-data fusion (MDF) framework. The MDF framework implements DALEC-Grass probabilistically using location-specific (1) weather time-series, (2) soil data and (3) earth observation (EO) time-series on vegetation leaf area, in order to (A) calibrate the model's parameters (B) validate model predictions and (C) estimate field-scale C cycling i.e. C in biomass, roots, soil, C lost to the atmosphere and C removed via grazing/cutting.
The package handles the sourcing and processing of all the necessary data as well as the implementation of the model's code. Its different functions allow the user to :
- Collect earth observation (EO) data from the ESA Sentinel-1 (SAR) and Sentinel-2 (multispectral) systems
- Process the EO data into weekly continuous time-series of grass Leaf Area Index (LAI)
- Implement the MDF algorithm
- Python 3
- A user account on the Alaska Satellite Facility archive centre
- A user account on the European Centre for Medium-Range Weather Forecasts
- A user account on Amazon Web Services
- The Sentinel Application Platform should be installed on your system
To install either "git clone https://github.com/vmyrgiotis/MDF_DALEC_GRASS.git" and then run "python setup.py MDF_DALEC_GRASS/install" --OR-- "pip install git+https://github.com/vmyrgiotis/MDF_DALEC_GRASS.git"
To run the tutorial on a local machine :
- install miniconda on the machine (if not installed) : https://docs.conda.io/en/latest/miniconda.html
- create a conda environment by running : "conda create -n dalec_grass python=3.9.7"
- activate the dalec_grass conda environment : "conda activate dalec_grass"
- install some python packages : "pip install matplotlib jupyter pandas numpy spotpy netCDF4 wand itermplot salem convertbng geopandas motionless joblib xarray"
- clone the github repo : "git clone https://github.com/vmyrgiotis/MDF_DALEC_GRASS.git"
- navigate to the location of the github repo that you just cloned : "cd /MDF_DALEC_GRASS"
- run "jupyter notebook" , your default browser starts , click on dalec_grass_tutorial.ipynb
To run the tutorial on a remote server:
- open a 1st terminal tab and : "ssh [email protected]"
- complete steps 1 to 6 from the running-on-local-machine shown above
- then : "jupyter notebook --no-browser --port=8008"
- open a 2nd terminal tab and : "ssh -L 8008:localhost:8008 [email protected]"
- on the local machine open an internet browser and go to http://localhost:8008/
To clean up after running the tutorial
- conda deactivate
- conda env remove -n dalec_grass
- rm -r MDF_DALEC_GRASS