To what degree does historical measurement inaccuracies contribute to climate change models?

Question

Often the graphs depicting the earth's historical temperature, the time axis extends well over couple of centuries. Also there are few that extends over thousands of years.

I wonder to what degree the measurement inaccuracies are taken into account while using these charts to communicate the ongoing climate change. Its a valid question that older temperature measurement suffer a higher error compared to modern day's methods.

Even the feared 2.5°C temperature rise over the industrial revolution time is a minute variation over centuries. To what degree has the climate models taken into account these measurement errors into account? (In short, the question is basically from a climate change sceptic that goes like - "Thermometers have gotten better and we are just measuring the true temperature and it's not that the world is warming up. So how can we accept climate change/global warming?")

Answer 1

I would estimate 0%, as long as we are considering the same types of models. See, climate models and in the regard that you may think of observations. Climate models aren't really initialized on observations of the system, but rather on inputs/forcings and mean state.

Observations of today's climate may validate the accuracy/tuneness of a climate model, but the types of observations used to run climate models may be different from what you expect. For example, the thermometer was invented either near 10 AD or in the 17th century, so it was obviously not a widespread invention then. Usually, those temperature measurements are taken as proxies, such as the amount of oxygen isotope or atmospheric composition in ice cores. So the answer to the climate skeptic is that it isn't just the temperature itself, but the modern day observations of temperature proxies that show past warming.

Answer 2

"In short, the question is basically from a climate change sceptic that goes like - "Thermometers have gotten better and we are just measuring the true temperature and it's not that the world is warming up. So how can we accept climate change/global warming?"

You could point them to the satellite datasets, which show that there has been significant warming since the late 1970s.

(source)

You could point out that instrumental record has error bars that are published with the data that reflect the decrease in uncertainty over time, e.g.

which makes it clear that the warming we have observed cannot be explained by the uncertainty. The best group provide code and papers allowing any skeptic to check their methodology (although I can't find the link just at the moment).

Similarly, the uncertainties of paleoclimate proxy reconstructions are not able to explain the current warming either (especially the rate at which it is happening).

Often the diagrams shown in the media do not include the uncertainty, but generally the uncertainties are stated in the original data sources.

Answer 3

The graphs you mention, with temperatures extending for centuries or even millennia, are actually not measurements from thermometers. They are based on proxies such as tree rings or ice cores, from which temperature is estimated, not measured. Of course this estimate is based on some measurable variable (tree ring width, for example), which itself has an uncertainty, but this uncertainty does not vary with time. If you analyse tree rings from a 1000 years old tree, measurements of width (or wood density) will be as accurate on the oldest rings than on the more recent ones, as all will be accomplished with the same, modern method.