La lien dans vote profile indique que vous avez travailler avec la modèle IMACLIM et l'IIASA. Donc vous avez beaucoup d'éxperience dans cette domain.
Admittedly, this is the phase 5 documentation, whereas the graph uses phase 4 results.
See the technical documentation document available here: www.ngfs.net/en/publicati...
Regarding a recent community science paper on 1.5°C of #globalwarming (doi.org/10.5194/essd...), the graph I like the most shows #seaice extent anomalies for different datasets and reconstructions. The most striking feature is HadISST2 disagreement with Antarctic reconstructions.
I intended for the graph to be accessible to red-green colorblind people, but authors kept finding additional recent reconstructions. Here is my attempt at a red-green accessible version. Feedback would be welcome. The community paper is open for public comment until March 11.
However, many other datasets might underestimate #warming despite using HadISST2 since they often don’t account for climatological differences between #seaice and open sea. This issue is discussed in section 3 of the paper as well as in my 2024 paper (doi.org/10.1002/qj.4...).
As a result, two datasets that I produce (HadCRU_MLE and DCENT_MLE), and perhaps Berkeley Earth, overestimate #warming since 1850. My 2024 paper (doi.org/10.1002/qj.4...) describes another mechanism (discontinuous temperature fields) that could contribute to overestimation.
The 1929-1939 climatology was designed to be conservative to warn sailors of potential #seaice danger rather than to be representative of average #seaice conditions. This is relevant since almost all global instrumental temperature datasets since 1850 rely on HadISST2. doi.org/10.14430/arc...
Prior to 1940, HadISST2 uses a German 1929-1939 reconstruction, which included observations from whaling ships, the British-Australian-NZ Expedition, and Nazi attempts to claim New Swabia for the 3rd Reich prior to WW2. This climatology was likely misinterpreted.
Regarding a recent community science paper on 1.5°C of #globalwarming (doi.org/10.5194/essd...), the graph I like the most shows #seaice extent anomalies for different datasets and reconstructions. The most striking feature is HadISST2 disagreement with Antarctic reconstructions.
Two other comments.
1. Mitigation costs in the graph are largely the result of a global uniform carbon tax. If society chooses a less optimal policy then mitigation costs can be much higher.
2. At least the graph doesn't exaggerate damages via a 8.5 W/m2 scenario.
The source is the Network for Greening the Financial System. Their results and methodological details are all available from their website. Unfortunately, the graph has many issues, such as using a damage function that has been retracted by nature.
There are more reliable estimates of climate damages that can still communicate that climate change can have substantial negative impacts. Please use alternative estimates. The graph you provide is unfortunately misinformation about climate change.
The Kotz et al. damage function conceptually includes extreme weather effects, so even if there were no Uzbekistan issue, this graph double counts damages. Finally, the NGFS acute damage estimates are biases since they account for extreme heat changes but not extreme cold changes.
NGFS phase 4 used the kotz et al damage function, which has been retracted from nature due to its results being based on faulty Uzbekistan data. www.nature.com/articles/s41... Futhermore, you graph adds NGFS acute and chronic damages.
How well can we quantify when 1.5 °C of global warming has been exceeded?
Open review until 11 March
essd.copernicus.org/preprints/es...
The paper describing the new DCENT-I dataset of monthly global surface temperature since 1850 is just published.
rmets.onlinelibrary.wiley.com/doi/10.1002/...
Statistical forecasts can be okay, but people proposing such models need to check for overfitting using standard tests. If there is physical information available then it makes sense to try to account for that.
William Nordhaus (Nobel prize winner) proposes a carbon compact where a coalition of countries agree to a carbon tax rate and tariff non-compliant countries. This would be a more reliable way to deal with the free rider problem of emissions reductions.
Clean energy targets are less economically efficient compared to broad-based carbon taxes as the latter treats all emissions reductions equally so better allows for competition between emission reductions. You can also relate carbon tax rates to the social cost of carbon.
HILL-1 appears to be based on RCP8.5, including 95th percentile results. It is surprising that this is described as "plausible". I thought that RCP8.5 was no longer considered to be part of the range of plausibility, including by the working group developing CMIP7 scenarios.
Let's add some nuance here. The direction of impact that AI will have on #climate emissions is disputed: www.nature.com/articles/s44...
The Kotz et al. paper is also quite relevant as it was ranked the #2 #climate paper of 2024 (www.carbonbrief.org/analysis-the...). The Kotz et al. damage function has also been incorporated into REMIND, the world's most used process-based integrated assessment model. While the CMIP7 scenarios...
... will likely use REMIND for some scenarios, similar to previous CMIP cycles, the CMIP7 scenario guidelines instructed IAM groups to neglect climate damages from simulations. Hopefully by the 2030s, damage function issues will be sufficiently resolved for inclusion in CMIP8 scenarios.
The Kotz et al. paper is also quite relevant as it was ranked the #2 #climate paper of 2024 (www.carbonbrief.org/analysis-the...). The Kotz et al. damage function has also been incorporated into REMIND, the world's most used process-based integrated assessment model. While the CMIP7 scenarios...
I have been following this issue for a few months. The Kotz et al. study unintentionally used faulty Uzbekistan data (e.g. some years with more than 100% gdp growth) which explained most of their results (www.nature.com/articles/s41...). Kotz et al. have since released an unpublished update...
that fixes the faulty Uzbekistan data and includes other improvements (www.pik-potsdam.de/en/news/late...). The issue is quite relevant since the Kotz et al. results were adopted by the Network for Greening the Financial System and thus various central banks and regulatory institutions.
I have been following this issue for a few months. The Kotz et al. study unintentionally used faulty Uzbekistan data (e.g. some years with more than 100% gdp growth) which explained most of their results (www.nature.com/articles/s41...). Kotz et al. have since released an unpublished update...
The form of the argument. You can be wrong due to a incorrect premise rather than due to a fallacious argument.
I'm no expert on this or its nuanced tradeoffs. If you have sufficient computational power and data, then wouldn't earlier better?
