Numerical Weather Prediction on AWS & AWS-Graviton2 with WRF

Launch an EC2 instance running WRF (advanced options)

  1. NICE-DCV ( is available for high-performance display and preinstalled in the AMI. However, and unlike for clusters, you will need to manually start the server ($ sudo systemctl start dcvserver) and create the session ($ dcv create-session wrfdisplay). The session will be available in most web browsers via https://External_IP_instance:8443/#wrfdisplay.
  2. The subdirectory /home/centos/DATA contains several preprocessing data of interest:
    2.1. Preprocessing data for the January 2000 example case (
    2.2. A file detailing the content of the CESM LENS (Community Earth System Model Large Ensemble). Files can be downloaded with the command aws s3 (e.g. aws s3 cp s3://ncar-cesm-lens//filename). Add the flag ‘–no-content’ if you still have not configured your AWS CLI credentials.
    2.3. A file OpenData_list with the endpoints of available datasets. Other datasets might become available in the future, and we will update the list accordingly.
  3. NCL (NCAR Command Language) is available for post-processing duties. The official website ( furnishes documentation on the language itself and how to create high-quality graphics from WRF data.