
This page shows “GMST Milestones Anomalies”, using purely Copernicus ERA-5 data (see GMST Data Sets). As per Climate Reporting – Why so many different values, the Copernicus ERA-5 data tends to be midway between NOAA/GISSTemp (low end) and Berkeley Earth (high end). The Copernicus ERA-5 data shows acceleration a bit more sharply than other data sets.
The trend line is calculated using Loess smoothing, with a 30-year window.
There is a similar graph at Climate Milestones – NOAA, GISS, Copernicus, HadCRUT, Berkeley Earth showing the same information but using a mixed GMST Data Sets.
The graphic above (using Copernicus-only-data), showsthe “GMST Trend °C” passing through “GMST Anomaly milestones” slightly earlier than with the Mixed data sets.
- 1985 – “GMST Anomaly Loess Trendline” passes 0.50°C
- 2000 – “GMST Anomaly Loess Trendline” passes 0.75°C
- 2011 – “GMST Anomaly Loess Trendline ” passes 1.00°C
- 2020 – “GMST Anomaly Loess Trendline” passes 1.25°C
- 2025 – “GMST Anomaly Loess Trendline” reaches 1.40°C in Feb 2025.
- Note that as per Trends and Projections, the Loess trend line does move as new data is added.
Trend lines weave through the Yearly averages

Following my twitter post of the 1000 Days above 1.5C before Trend passes 1.5C? graphic, I received some excellent criticism, I made a few updates, to check the trend lines. Note. I am planning to post updates to the graphics every month or so, and I will refine the designs as I get useful feedback.
- The graphic above shows the individual yearly averages, so you can see the trend lines weaving through the individual years. 2021 / 2022 were below trend (and part of triple dip La Nina).
- The trend line now uses Loess-smoothing with a 30-year-window. This is an improvement on my homemade smoothing line I used previously
- The trend lines go above other ones I’m seeing posted by establishment figures / authorities, so it will be interested to see what trend algorithm I can find that produces those trend lines.
- I have stopped calling it “The Climate”, and started calling it “Loess 30-year Trendline”, to make the graphic more objective.
- Previously my graphic stated “The Climate” passed some value, and that was overreach on my part, and probably devalues / distracts-from the graphics. I have moved to loess-smoothing, and then I can just say the loess trend line reached x-C in year y.
- There seems to be an interesting debate between:
- “Michael Mann Analysis”:
- 2023 / 2024 Data / Observations are in keeping with CMIP6 models
- Acceleration has not been proven. “The truth is bad enough”
- 2025: Again MM points to a paper that says statistically any acceleration is not proven (yet).
- “Hansen Analysis”:
- Keep an eye on the 2024 data: Acceleration is highly plausible
- Hansen has significant concerns that the models have the GHG / Aerosol / Climate Sensitivity incorrect.
- Hansen points to paleo-climate data and the IMO “inadvertent sulphur experiment”
- Hansen’s expectations are more in line with my trend lines, which is interesting but of no practical importance.
- I have no scientific qualifications, so will stick to my lane of making interesting graphics, and make it clear how I produce my them. Any projections I might play with, will be projections suggested by others, and I will just show my understanding of the implications of those projections.
- It will be super interesting to see how the trends play out over the next years.
- I’m going to play with the different graphs produced by different entities, so I can understand more, what the differences are.
- “Michael Mann Analysis”:
- There seems to be an interesting debate between:
How the Graphic was Created
Same technique as outlined in Climate Milestones – NOAA, GISS, Copernicus, HadCRUT, Berkeley Earth … but just using Copernicus data.