Data Tampering of Marcott by Javier Vinós

The graph above (and similar versions of it) is circulating widely on social media and promoted as evidence that the Holocene Thermal Maximum (HTM) between 5000 and 10,000 years ago was warmer than today. The graph originates from Javier Vinós' self-published book. In the graph above the black "b" time series is reported to be variability of global temperatures since the beginning of the Holocene. This particular version also shows Marcott et al 2013 as the "a" time series and Earth's obliquity as the "c" time series. My interest here has to do with the data tampering Vinós used to generate the black "b" time series, so I'm going to refer to this as the "Vinós time series."

The Vinós time series is fabricated by tampering with Marcott's proxy data. In his self-published book and various blogposts, he has told us how he fabricated it. To see how he did so, we must first look at a version of this graph in which the Vinós time series is plotted as Z-score instead of temperature.

Beginning with the graph immediately above, here's how he constructed the graph at the top of the post. I've linked to the sources he's claimed to have used as well as the posts where he describes what he has done so you can check up on me to make sure I'm describing his process accurately.

1. Vinós used a "differencing method" to average the proxy data from Marcott, a method that he claims follows a blogpost by Tamino, in an attempt to account for proxy drop off at the end of the time series (essentially the first half of the 20th century). Vinós calculated Z-scores instead of temperatures here, which will become significant below.
2. Vinós claims to have used Tamino's method to correct for proxy drop off and an artificially enhanced warming spike in the 20th century. The latter in part in particular is discussed by Marcott in his paper, but Vinós didn't mention that Marcott's RegEM (area-weighted) reconstruction already accounts for proxy drop off and produces 20th century warming consistent with the instrumental record. The discussion in Tamino's post (which I discuss here) had to do with whether the difference between Marcott's Standard 5x5 and RegEM reconstructions for first half of the 20th century was robust. Tamino argued that it would have been robust if Marcott used the differencing method for the Standard 5x5.
3. Vinós differs from Tamino in that he plotted the Vinós time series as Z-scores instead of temperature. He then scaled his time series differently from and Marcott's Standard 5x5 reconstruction to make the HTM appear warmer in Vinós' time series than in Marcott's.
   
4. In the version at the very top of this post Vinós plotted his time series as temperature rather than as a Z-score, and now in the Vinós time series shows the difference between the HTM and Little Ice Age (LIA) to be about 2x what is shown in the proxy data from Marcott. On Judith Curry's blog, Vinós explained why he tampered with data by inflating the temperature scale. According to him, the "temperature anomaly was rescaled to match biological, glaciological, and marine sedimentary evidence, indicating the Holocene Climate Optimum was about 1.2°C warmer than LIA." It's actually about 1.4°C in his graph.
5. His explanation is dishonest because the proxies used by Marcott already included "biological, glaciological and marine sedimentary evidence" (see map below), and that evidence showed that the HTM was only about 0.7°C warmer than the LIA. Vinós did not rescale his graph to match this evidence but to deviate from it. He just doubled the difference between the LIA and HTM by manipulating the temperature scale in his time series. In fact, you can quantify exactly the amount of data tampering that Vinós used if you compare the Vinós time series to the Marcott time series in the top graph above. 

Map of "biological, glaciological, and marine sedimentary evidence"
used in Marcott 2013, from Figure S1

Vinós tries to justify his data tampering by saying that there is "vast literature" on his side. He says, "I have also rescaled the temperature changes to make them congruent with the vast literature and consilience of evidence from different fields that indicates that the Holocene Climatic Optimum was on average between 1 and 2 °C warmer than the Little Ice Age." So Vinós read somewhere that the HTM should be globally 1°C to 2°C warmer than the LIA, and because the actual temperature data didn't match what he believed it should be, he just changed the data to match his expectations. And where is the "vast literature" showing that the HTM was globally 1°C to 2°C warmer than the LIA ? He predictably doesn't say. Vinós essentially admits to data tampering here. The proxy data didn't match what he believed it should show, so he changed the data to match his expectations. The Vinós-tampered version doubles the difference between the HTM and the LIA, changing that difference from 0.7°C to 1.4°C.

A Correct Plot of Marcott 2013

Above is a correct plot of Marcott et al 2013 using his RegEM method that already accounts for proxy drop of near the end of the time series (so that the the 20th century warming trend isn't exaggerated) along with HadCRUT5 set to the same smoothing as is in Marcott's proxy data. 

I've seen several versions of the Vinós time series posted on social media and political tracts, including Curry's and Andy May's blogs and even in a manuscript by Nicola Scafetta. Sometimes the Vinós time series is titled with Vinós' name and sometimes with Marcott's name, a trick that can make it more difficult to know if you're looking at Marcott's reconstruction or the Vinós-tampered version. The easiest way to tell is to look at the difference between the LIA and HTM; if the difference is ~0.7°C it's probably legit. If it's >1°C, it's probably the result of Vinós' data tampering.

References:

[1] Bova, S., Rosenthal, Y., Liu, Z. et al. Seasonal origin of the thermal maxima at the Holocene and the last interglacial. Nature 589, 548–553 (2021).
https://doi.org/10.1038/s41586-020-03155-x

[2] Osman, M.B., Tierney, J.E., Zhu, J. et al. Globally resolved surface temperatures since the Last Glacial Maximum. Nature 599, 239–244 (2021). https://doi.org/10.1038/s41586-021-03984-4

[3] Marcott, Shaun et al. “A Reconstruction of Regional and Global Temperature for the Past 11,300 Years.” Science 339 (2013): 1198-1201. http://shpud.com/Science-2013-Marcott-1198-201.pdf