Public-Facing Tool Summarizing Local CMIP6 Projections

CMIP6 Hackathon CMIP6 Project Proposals
1

Public-Facing Tool Summarizing Local CMIP6 Projections

Scientific Motivation

Decision makers and the general public are often most interested in what cutting-edge climate projections mean for a specific location. However, despite CMIP model output being publicly available, the technical skills required to process, view, and understand climate model output prevents nontechnical audiences from accessing this information. While governments and nonprofits eventually analyze location-specific climate projections in climate mitigation and adaptation reports, such reports can take years to develop making the provision of local climate projections out of sync with popular press surrounding advancements in the latest generation of climate models. Additionally, these reports often gloss over information regarding model or scenario uncertainty. While excluding such detail can help simplify the results, they often leave audiences unclear about how the process of climate modeling works, including confusion over what creates the uncertainty in such climate modeling output.

We propose to develop a tool that addresses this need while also serving as an outreach tool to help the public better understand climate models and uncertainty. In particular, we aim for the tool to improve communication of the following concepts:

  • The local climate consequences of global socioeconomic decisions
  • The difference between scenario and model uncertainty
  • The difference between global and local changes (e.g. what does a 2°C world mean for temperatures in Seattle?)

Proposed Hacking

We propose to develop a public-facing interactive tool that enables users to input latitude and longitude coordinates, or major cities, and see a dashboard of figures showing location-specific information about historical and future temperature and precipitation trends for that location. We currently plan to use raw CMIP6 model output without any statistical downscaling or bias correction, but in the future the tool could be extended to include these modifications or historical observations, or to compare CMIP5 and CMIP6 model output. Example figures include:

  • Time series of multi-model mean local temperature changes for historical record and various SSPs
  • Time series of mean local temperature changes for various CMIP6 models for a given SSP
  • Box plots showing the projected local temperature change at different levels of global warming (1.5°C, 2°C, 3°C, and 4°C worlds)

We propose to develop this tool in three phases:

  1. Develop schematic of tool, including options for interactive user input and figures to meet outreach objectives
  2. Code interactive offline tool in Python
  3. Bring tool online to be public facing

We aim to complete at least the first two phases during the hackathon.

Anticipated Data Needs

Monthly-average fields from CMIP6 historical and future warming scenarios:

  • Near-surface air temperature
  • Average precipitation
  • Average precipitation as snow

Anticipated Software Tools

  • Standard Python / xarray
  • Heroku or other platform to deploy public-facing Python app
3 Likes
2

I think this is an exciting and creative idea–definitely something different from many of the other proposals!

This sort of application would fit very well with the google-cloud-based CMIP6 data archive, since it those data will be available for the app to compute anything it wants on the fly. We could consider creating some copies of certain variables that are optimized for the sort of timeseries analysis you have in mind here.

Key to making rapid progress at the hackathon will be effectively reusing existing tools that can make your life easier. Within Pangeo, we have been playing around with a few new tools in the interactive data viz space. Some of these are.

1 Like
3

Hi @czarakas! This is a great proposal and one that will certainly garner a fair bit of attention from those interested in interactive dashboards and those interested in producing policy relevant data visualizations. You may consider reproducing a few key figures from the AR5 Summary for Policy Makers or Executive Summary (e.g. https://www.ipcc.ch/report/ar5/syr/summary-for-policymakers/) report, except with CMIP6 data. I think would this would provide a nice entrypoint to an interactive tool where users could dig in further.

While I’m at it, some other interesting interactive vizualization tools that came out during the CMIP5 era: