Location: Budinich Lecture Hall

Conveners: Huikyo Lee (Jet Propulsion Laboratory, California Institute of Technology, USA), Rasmus Benestad (Norwegian Meteorological Institute (NMI), Norway), Abdelkader Mezghani (NMI), Kajsa Parding (NMI), Ian Brosnan (NASA Ames Research Center, USA), Alexander
Goodman (Jet Propulsion Laboratory, California Institute of Technology, USA).

This session will aim at training the users on how to perform a comprehensive analysis comparing both bias-corrected regional climate model output and empirical-statistical downscaling results against observations to obtain a robust and reliable climate change prediction. It will follow up work carried out during the #ConCordOslo20221 workshop and previous CORDEX D sessions.

We will explore CORDEX RCM output and statistically downscaled products, such as NASA Earth
Exchange Global Daily (NEX-GDDP-CMIP6) and region-specific ESD-seasonal Downscaled Projections. We will be using tools such as the R-esd package developed by the Norwegian Meteorological institute and the Regional Climate Model Evaluation System (RCMES) developed by NASA. This includes hands-on training about extracting, processing, downscaling, bias-correcting, visualizing, analyzing and manipulating datasets. The training will also address two approaches in ESD: “downscaling weather” and “downscaling climate” and we will explain new concepts such as “common EOFs” as a framework for ESD predictors.

The expected outcome of this training workshop is 1) to disseminate statistically downscaled multi-model CMIP6 GCM ensembles and open-source software to analyze the downscaled dataset to the CORDEX community, 2) to train potential users of NEX-GDDP-CMIP6 to make use of cloud-based resources, and 3) to increase communication and collaboration between different CORDEX domains.

Participants have to bring their own laptops to attend this training session.

Our training materials are hosted on our JupyterHub server, which can be accessed at the following link: https://ocw-jpl.org/

Participants can find their respective IDs for logging into the JupyterHub server from the email sent out yesterday. The password for all participants is “aist-ocw.”

Before the session starts (between 14:15-14:30), we will assign IDs to those who have not registered for D7 at Budinich Lecture Hall and Microsoft Teams.

We kindly request that all attendees who have registered for our D7 session visit https://ocw-jpl.org/ and test the system in advance.

Post-session survey link: https://www.surveymonkey.com/r/Z5VXSNN

Detailed program

14.30 – 15.30 (Speaker: Abdelkader Mezghani & Rasmus Benestad)
14.30 – 14.35 Welcoming words and introducing the session
14.35 – 14.45 Introducing Norwegian Meteorological Institute’s (NMI’s) Empirical Statistical Downscaling (ESD) methodology
14.45 – 15.30 Hands-on training (ESD)
1. Bring in your own observation data or use ERA5 or equivalent gridded data as observations for those who wants to downscale GCM ensembles using ESD,
2. Run sample/custom scripts to downscale precipitation and temperature
3. Downscale precip or temp to a region of interest (using ESD), or use bias-corrected data from NASA, visualize the results, make summary plots
4. Compare the downscaled results from a regional climate model (RCM) and ESD

15.30 –16.00 Health break (Q&A and troubleshooting continue)

16.00 – 16.55 (Speaker: Bridget Thrasher & Hugo Lee)
• 16.00 – 16.10 Introducing NASA’s statistically downscaled product (NEX-GDDP-CMIP6) and RCMES system
• 16.10 – 16:55 Hands-on training (NEX-GDDP-CMIP6)
1. Run RCMES commands and analyzing NEX-GDDP-CMIP6
2. Extracting a subregion of interest and manipulate the data
3. Process the results and produce summary plots and maps

16:55 – 17:00
• Collecting feedback from participants

Due to unforeseen circumstances some speakers may be replaced or will not be able to participate.

Contact: Huikyo Lee (huikyo.lee@jpl.nasa.gov), Rasmus Benestad (rasmusb@met.no), Abdelkader Mezghani (abdelkaderm@met.no). Kajsa Parding (kajsamp@met.no), Ian Brosnan (ian.g.brosnan@nasa.gov), Alexander Goodman (alexander.goodman@jpl.nasa.gov).