Climate Change Visualizations

What are the latest projections of climate change?

How likely are they?

Let's visualize the latest knowledge of all CMIP6 models and their variability.

Global temperature projections

The widest selection of models covering the main scenarios where emissions won't grow

Models covering also 1.5 °C scenario (ssp119) of the Paris agreement

This scenario now seems out of our reach.

Context

Averages by the 50th quantile. Ranges by the 10-90th quantile. Normalized to the last 20 years of hindcast.

Models availability (Aug '24) for the whole period:

18 models for all 3 scenarios: CAMS-CSM1-0, CanESM5, CanESM5-1, CNRM-ESM2-1, EC-Earth3, EC-Earth3-Veg, EC-Earth3-Veg-LR, FGOALS-g3, GFDL-ESM4, GISS-E2-1-G, GISS-E2-1-H, IPSL-CM6A-LR, MIROC-ES2H, MIROC-ES2L, MIROC6, MPI-ESM1-2-LR, MRI-ESM2-0, UKESM1-0-LL

47 models for ssp126 and ssp245: Previous 18 + ACCESS-CM2, ACCESS-ESM1-5, AWI-CM-1-1-MR, BCC-CSM2-MR, CanESM5-CanOE, CAS-ESM2-0, CESM2, CESM2-WACCM, CIESM, CMCC-CM2-SR5, CMCC-ESM2, CNRM-CM6-1, CNRM-CM6-1-HR, FGOALS-f3-L, FIO-ESM-2-0, GISS-E2-1-G-CC, GISS-E2-2-G, HadGEM3-GC31-LL, IITM-ESM, INM-CM4-8, INM-CM5-0, KACE-1-0-G, KIOST-ESM, MCM-UA-1-0, MPI-ESM1-2-HR, NESM3, NorESM2-LM, NorESM2-MM, TaiESM1

48 models for at least one of the scenarios: Previous 47 + EC-Earth3-CC

Variance in climate models

To interpret the variance among the CMIP models, the Intergovernmental Panel on Climate Change (IPCC)_ assesses the accuracy of the models by applying statistics to determine which model projections are more consistent with lines of evidence. They combined models by weighting them by their likelihood into an ‘assessed warming’ projection in their Sixth Assessment Report (AR6).

Currently, the biggest debate is about so-called 'hot models' that predict the highest warming. In the IPCC 'assessed warming', they are considered to be less likely, because they less accurately reproduce historical temperatures over time.

Alternatively, a subset of models with 'TCR' between 1.4-2.2 ºC can be used to produce very close results, as described in the Nature.com article. Nature, is one of the top scientific journals recognized for its rigorous peer-review process involving multiple rounds of evaluation by experts in the field.

• TCR: Transient Climate Response is the amount of global warming in the year in which atmospheric CO2 concentrations have finally doubled after having steadily increased by 1% every year.

• IPCC: Intergovernmental Panel on Climate Change is an international body established by the United Nations Environment Programme (UNEP) and the World Meteorological Organization (WMO) to assess the scientific, technical, and socio-economic information relevant to understanding the risk of human-induced climate change.

• • The IPCC's assessments are based on the work of thousands of scientists from around the world and are widely regarded as the most comprehensive source of information on climate change. The IPCC does not conduct its own research but synthesizes existing scientific literature, including that made by CMIP. Its reports go through a rigorous process of review and revision involving experts to ensure accuracy, transparency. They provide policymakers with regular scientific assessments of the current state of knowledge about climate change.

Future Scenarios

The largest source of uncertainty in global temperatures in decades to come is the volume of future greenhouse-gas emissions, which are largely under human control.

The scenarios visualized:

1. Reaching worldwide carbon neutrality around 2050 => projected 1.5° global warming – probably out of our reach (ssp119)
2. Carbon neutrality around 2075 => projected 2° global warming – will we make it? (ssp126)
3. No decline of emissions until 2050 => projected 3° global warming (ssp245)

There are even more pessimistic scenarios like "Double emissions in 2100 => projected 4° global warming (ssp570)" that could be visualized.

Maximum temperature projections (WIP)

Maximum temperature in Czechia

Tropic days annualy in Czechia

Context: Work In Progress

This is an early discovery. The max temperature aggregation is yet to be reviewed. It seems that the model resolution lowers the maximum temperatures a bit and more complex approach would be needed for more accurate values, not just the trend.

For example a CLIMRISK uses a statistical interpolation and avaluates model historical accuracy.

Where are the projections from?

The most widely recognized projections are from the Coupled Model Intercomparison Project, phase 6 (CMIP6), run by the World Climate Research Programme.

• CMIP is a collaborative framework designed to improve our understanding of climate change through the systematic comparison of climate models. Their work is publicly shared with anyone for use or review.

• CMIP shares around 50 distinct climate models (Aug 2024) created by 49 modelling groups from 17 countries all across the world from Europe, America, Asia, Australia and Oceania.

Data Sources

Model data from Copernicus Climate Data Store or Earth System Grid Federation Data Portal.

Measurement data from Met Office and Czech Hydrometeorological Institute

How to run it

1. install python libraries in requirements.txt
2. Register on Data Store:
   1. Copernicus (https://cds.climate.copernicus.eu/):
      • get an API key in your user profile
      • put your API key to ~/.cdsapirc as in this format
   2. ESGF (https://esgf.github.io/nodes.html) – any node:
      • put the open-id you'll get after the registration to ESGF_OPENID=*** & ESGF_PASSWORD=*** lines into ~/.esgfenv
3. change DATADIR to the folder point where to download model data
   • Toggle DownloaderCopernicus/DownloaderESGF to get missing models on either of them
4. Run python3 ClimateProjections.py
   • preview to skip download and visualize already downloaded data
   • max to visualize the local maximum temperature instead of the global average
   • esgf to download from the ESGF instead of Copernicus data store. wget to use their script instead the direct download.
   • Update model experiment in the code when needed.

Tested only on python3.

Downloading

Downloading is tricky, because not all models are available on all servers. Multiple download methods might be necessary to use when the method is unable to download a particular model.

• DownloaderCopernicus is easier to start with, but more models are missing there.
• DownloaderESGF has more models, but can be hard to tame (run with esgf parameter). It might be necessary to add trusted certificates (an error suggests how) or try using their scripts with some dependencies (run with wget parameter).

Also, sometimes a server timeouts, so it might be necessary to wait hours before retrying. Somebody is welcome to contribute by making it more robust. Plus at this point the ESGF python library does not limit data by geographic coordinates, which makes it cumbersome for huge daily models. The program visualizes just the aggregations starting with agg*, so the big source files can be deleted once aggregated.

If you break the download in the process, remove unfinished files not to have holes in the data. One serie is sometimes split into multiple files, so all files with the last ssp experiment and model names are recommended to be deleted in such case to make downloader redownload it. Watch for yellow warnings of duplicates or incomplete data.

All models have tens of Gigs. It takes time.