One of the many misleading things done by climate scientists is to splice together data derived via different measurement methods.
The CO2 curve 800000 years back is an excellent example.
CO2 in air bubbles in ice will diffuse into the ice during the many millenia the ice have been stored under pressure. You can therefore expect the ice core CO2 to be lower than hypothetical atmospheric measurements at the same time.
The diffusion is expected to progress fastest at the start, and then more slowly.
Still the artifact is absolutely obvious as the CO2 concentration peaks are lower
and lower the further back in time it goes.
Yet of course someone had to splice it all together and not even add error bars.
Not sure if you are trolling of really believe this to be true, but for anyone else assuming good faith: The diffusion aspect is (of course) well-modeled, see e.g. references under "past greenhouse gasses" at [0].
They have found some meltwater layers with unexpectedly large quantities of CO2 > 750ppm.
But they also show that there is unexpected heavy diffusion around the meltwater layers. And they argue very compellingly, that the ice core CO2 records have been smoothed through natural diffusion.
Why is this information not given when showing this graph?
Where is it well-modelled? Noble gasses will diffuse differently from CO2. And on-top of diffusion of course CO2 will make hydrogen bonds with water in the ice.
It is well-understood that many ice cores give the same relative shape of peaks of various gasses through time.
Going from there to claiming knowing the absolue concentrations without very large error bars, is just not science.
"Carbon dioxide measurements from older ice in Greenland is less reliable, as meltwater layers have elevated carbon dioxide (CO2 is highly soluble in water)."
The CO2 curve 800000 years back is an excellent example.
CO2 in air bubbles in ice will diffuse into the ice during the many millenia the ice have been stored under pressure. You can therefore expect the ice core CO2 to be lower than hypothetical atmospheric measurements at the same time.
The diffusion is expected to progress fastest at the start, and then more slowly.
Still the artifact is absolutely obvious as the CO2 concentration peaks are lower and lower the further back in time it goes.
Yet of course someone had to splice it all together and not even add error bars.