What is the margin of error for the temperatures in your chart?
Direct combination of these interpreted geological temperature records is not necessarily valid, nor is their combination with other more recent temperature records, which may use different definitions. Nevertheless, an overall perspective is useful even when imprecise. In this view time is plotted backwards from the present, taken as 2015 CE. It is scaled linear in five separate segments, expanding by about an order of magnitude at each vertical break. Temperatures in the left-hand panel are very approximate, and best viewed as a qualitative indication only.[9] Further information is given on the graph description page.
Data
Panel 1: 540 to 65 million years ago
The panel 1 data is from stable oxygen isotope measurements from the shells of macroscopic marine organisms, collected by Veizer et al (1999),[1] as re-interpreted by Royer et al (2004).[2] The graph effectively reproduces the upper panel of Royer et al's figure 4, but with an expanded range (see below). The orange band shows the effect of extreme assumptions in application of the GEOCARB reconstruction to interpretation, and is not representative of the full uncertainly (which would be much larger).
Because the Royer and Veizer results are indicative of the temperature of the shallow tropical and subtropical seas where the organisms lived,[2] they are unlikely to be fully representative of global average surface air temperature variation. The anomalies are plotted here expanded by a factor of two, as a very approximate conversion. Multiple confounding factors affect interpretation of samples this old, so panel 1 is best viewed as a qualitative indication of temperature (warmer/colder).[3]
Panel 2: 65 to 5.3 million years ago
This data is from the Hansen et al (2013)[4] interpretation of the global collection of oxygen isotope data from microscopic marine organisms of Zachos et al (2008).[5]
This is a direct estimate of global average sea surface temperature, a close analogue of surface air temperature. Hansen et al describe it as a "first estimate", meaning an approximate one, but limited independent corroboration (e.g. Zachos et al (2006)[6] for the Eocene optimum) indicates that it is substantially more quantitative than panel 1.
Panel 3: 5.3 to 1 million years ago
This data is from the Lisiecki and Raymo (2005)[7][8] global stack of oxygen isotope data from microscopic marine organisms interpreted using the Hansen et al (2013)[4] prescription.
At this scale, the Zachos et al stack (which also covers this interval) is virtually indistinguishable from the Lisiecki and Raymo stack. This is a direct estimate of global average sea surface temperature.
Panel 4: 1 million to 20,000 years ago
Two datasets are plotted:
Lisiecki and Raymo, as in panel 3.
Temperature estimates from stable hydrogen isotope measurements from the EPICA Dome C ice core from central Antarctica[9] These temperature anomaly estimates are polar, not global, and are here divided by a standard polar amplification factor (2.0, as for example in Hansen et al (2013)[4]) to approximately convert them to global estimates.
Panel 5: 20,000 years ago to present (2015)
Five datasets are plotted:
EPICA Dome C, as in panel 4.
Temperature estimates from oxygen isotope measurements on the north Greenland ice core, NGRIP,[10] interpreted using the simple procedure of Johnsen et al (1989).[11] (There are more modern and complex procedures which would yield slightly different interpretations.) Like the EPICA Dome C record, this record is polar, and is shown divided by a polar amplification factor of 2.0. The difference between this and dataset 1. illustrates the polar sea-saw hypothesis.
Global temperature estimates over the ~12,000 years of the Holocene from the multi-proxy collection and interpretation of Marcott et al (2013).[12]
Instrumental (not proxy) data since 1850 from the Berkeley Earth project land-ocean dataset (2014),[13] plotted as decadal means.
Projected temperatures for 2050 and 2100 from the IPCC Fifth Assessment Report's WG1 Summary for Policy Makers (2013)[14] for the RCP8.5 scenario.
See section at bottom:
https://en.wikipedia.org/wiki/File:All_palaeotemps.png#Summaryfrom the original link that shows this and other charts:
https://en.wikipedia.org/wiki/Geologic_temperature_record