Scenarios

This repository contains analytical tools for CMIP6 models ranking according to their correlation with ERA5 data.
The comparison works either at level of climatically relevant areas (close to federal districts) or at the level of administrative division of Russian Federation.
It accounts 16 CMIP6 models and 4 Shared Socioeconomic Pathways in total.

Dependencies

Dependencies are listed in requirements.txt file.

Data

The data required for the project is available in the mnt\public-datasets\taniushkina\RCP_scenarios\data folder. It contains various files in .nc and .json formats.

CMIP files were loaded with the script and related context dictionary, where one should mention

• variable: current analysis uses tas - near-surface air temperature and pr - precipitation,
• source_id cooresponds to one of CMIP6 models,
• institution_id and experiment_id are redundant parameters.

Administrative boundaries were loaded from GADM project.
Districts division withing Russian Federation performed by hand and stored in .json files in data\districts folder.

The project directory structure should be organized as follows (tree depth is limited by 2):

.
|
|-- README.md
|-- data
|   |-- boundary
|   |-- cmip
|   |-- districts
|   `-- era
|-- data_processed
|   |-- 2015_2022
|   `-- yearly
|-- environments
|   |-- Dockerfile
|   `-- environment.yml
|-- notebook
|   `-- analysis.ipynb
|-- results
|   |-- csv
|   `-- pics
`-- src
    |-- correlation.py
    `-- process.py

Docker

To set up the project using Docker, follow these steps:

• Build the Docker image: docker build -t rcp .
• Run the Docker container:

docker run -it --name <CONTAINER NAME>
   -v <SOURCE DATA FOLDER>:/RCP/data
   -m 128000m --cpus=16 --gpus '"device=0,1"'
   --ipc=host
   -w="/RCP"
   rcp

As an example:

docker run -it --name RCP_cont
   -v /mnt/public-datasets/taniushkina/RCP_scenarios/data:/RCP/data
   -m 128000m --cpus=16 --gpus '"device=0,1"'
   --ipc=host
   -w="/RCP"
   rcp

Executing program

The program consists of several steps, all of them are displayed in the notebook/analysis.ipynb notebook.