Guest “damning with faint praise” by David Middleton
Kaufman, Mckay, Routson, et al., 2020
In my previous post, I noted a “funny pattern” in the latest Holocene climate reconstruction (Kaufman, McKay, Routson, et al., 2020). In this post, I will go into more detail as to why I chose the composite plus scale (CPS) method rather than the other four methods employed by the authors.
I think the paper represents a good faith effort to reconstruct global climate change over the past 12,000 years. The authors were very transparent about data and methods, even noting that no one really knows how to reconstruct climate at this sort of scale. While not a Mannian hockey stick or a Marcottian mess, they sill insisted on muting the low frequency component of the climate signal and directly comparing the high resolution instrumental record to the very low frequency, heavily smoothed multi-proxy reconstruction. That said, one of their five reconstruction methods did preserve the low frequency climate signal (yes, I am using the word “signal” correctly). This post will point out two things:
- Direct comparisons of the instrumental data to the reconstruction violate the basic principals of signal theory.
- Only the composite plus scale (CPS) method is consistent with the observations.
Signal theory violation, 15-yard penalty and loss of down
This is from the paper’s introduction:
The database is the most comprehensive global compilation of previously published Holocene proxy temperature time series currently available. It comprises a quality-controlled collection of high-resolution time series (average sample spacing of 164 years) with well-established time scales (average of 1.0 age control points per 1000 years) that was selected from a much larger collection of temperature-sensitive proxy records. The multi-proxy database includes a total of 1319 paleo-temperature records from 470 terrestrial and 209 marine sites where ecological, geochemical and biophysical proxy indicators have been used to infer past temperature changes. Among the variety of proxy types, alkenones and isotopes are the dominant sea-surface temperature proxies, whereas pollen and chironomids are the most common terrestrial temperature proxy types. Most of the records (97%) are available as quantitative temperature reconstructions calibrated to °C, whereas the remaining 42 records represent non-quantitative temperature-sensitive proxy records.Kaufman, D., McKay, N., Routson, C. et al., 2020
This is from “Timing and magnitude of peak Holocene global temperature”:
To bracket the likely range of the temporal resolution of the GMST reconstruction, we focus on intervals of 1000 and 200 years, and quantify the difference in their magnitude and timing of peak warmth (Fig. 4).Kaufman, D., McKay, N., Routson, C. et al., 2020
In my previous post, I treated the reconstruction as if it had a 100-yr resolution…
The resolution of the reconstruction is no better that 164 years… I should have treated it as if it had a 200-yr resolution. At a 200-yr resolution, HadCRUT4 looks like this:
While I still think the paper is a very good paleoclimatology effort, this paragraph is not supportable in any scientific manner.
The distribution of peak global temperatures during the Holocene can also be compared with recent temperatures. The GMST of the past decade (2011–2019) averaged 1 °C higher than 1850–190011. For 80% of the ensemble members, no 200-year interval during the past 12,000 years exceeded the warmth of the most recent decade. For the other 20% of the cases, which are primarily from the CPS reconstruction, at least one 200-year interval exceeded the recent decade. This comparison is conservative in context of temperatures projected for the rest of this century and beyond, which are very likely to exceed 1 °C above pre-industrial temperature12. Such projections place the temperature of the last decade into a long-term context that is more comparable with the Holocene GMST reconstruction. Furthermore, if the reconstruction is influenced by a Northern Hemisphere summer bias (discussed below), then the peak warmth would be overestimated and the recent warming would therefore stand out even more in comparison.Kaufman, D., McKay, N., Routson, C. et al., 2020
The reconstruction can’t resolve decadal scale temperature changes. At the resolution of the reconstruction, HadCRUT4 would be a single data point at 0.23 °C.
The Ice Age Goeth
This is what the authors had to say about the CPS reconstruction method…
Among the five reconstruction methods, CPS stands out prominently with its large temperature changes (Fig. 3), especially in the Northern Hemisphere (Figs. 1 and 2). For example, the median ensemble member of the CPS reconstruction shows that GMST warmed by about 3.9 °C between 12 and 10 ka compared to about 1.1 °C for the other methods. The median GMST during the period centered on 6 ka, the long-standing mid-Holocene target for paleoclimate modeling experiments (e.g., ref. 15) was 1.1 °C warmer than the 19th Century in the CPS reconstruction compared to about 0.4–0.5 °C for the other methods (Table 1).
Although it is an outlier, we do not have irrefutable evidence to exclude the CPS reconstruction, and cannot rule out the possibility that the other reconstruction methods underestimate the overall variance. The outcome of the CPS method depends on the validity of the target used for scaling, which is difficult to verify. The high amplitude of temperature changes reconstructed by CPS might reflect chronological and other uncertainties that average out century-scale temperature variance during the compositing, thereby increasing the relative magnitude of millennial-scale variance in the composite. When the composite is then scaled to the reconstructions of the past two millennia, which have more realistic century-scale variance, the millennial-scale variance (and thus the long-term trends) are artificially inflated. Nonetheless, as an independent approach, CPS contributes to a more complete sampling of the uncertainty space. We therefore retain CPS as one-fifth of the multi-method ensemble, and we focus on the median rather than the mean as the best representation of the ensemble central tendency. Excluding CPS from the ensemble does little to influence the median GMST reconstruction. For example, the mid-Holocene (6.5–5.5 ka) ensemble median is only 0.05 °C cooler when excluding the CPS members; namely, the five-method median is 0.51 °C (0.19, 1.35) versus 0.46 °C (0.17, 0.79) when excluding CPS members.Kaufman, D., McKay, N., Routson, C. et al., 2020
They grudgingly left CPS in the mix because: “Although it is an outlier, we do not have irrefutable evidence to exclude the CPS reconstruction, and cannot rule out the possibility that the other reconstruction methods underestimate the overall variance.”
However, they could have easily found evidence “that the other reconstruction methods underestimate the overall variance.” CPS is the only method that clearly resolved the Holocene Climatic Optimum, the Roman Warm Period, the Medieval Warm Period and Neoglaciation…
I generated a plot of CPS as an overlay of Figure 3 from the PAGES 12K paper and replaced the 2K inset with the historical climate periods and Neoglaciation from Grosjean et al., 2007 and a North America ice extent map from Dyke et al., 2003. I also included a 2σ confidence band from the 500 ensemble members. My version of the CPS mean is the dashed orange curve…
Does the North American ice extent at 12,000 calendar years ago look more like 1 °C cooler than 1800-1900? Or 4 °C cooler?
OK… geophysical inertia might explain how there could have still been that much ice with temperatures only 1 °C cooler than 1800-1900, but I don’t think it could explain this:
The other four methods show very little temperature change from 9,500 years ago up until 1850 AD… A period when we know that there was massive ice retreat over the first 5,000 years and ice advance (Neoglaciation) over most of the next 4,500 years. CPS is the only one of the four methods consistent with the Holocene evolution of ice sheets and glaciers in the Northern Hemisphere. It’s also consistent with the Holocene evolution of Arctic sea ice.
The CPS method is clearly consistent with the nearly ice-free conditions from 10,000 to 5,000 years ago and the neoglacial expansion of sea ice from 5,000 years ago to the mid-1800’s. The other four methods indicate very little temperature change over this time period.
What about Northern Hemisphere summer bias?
Three of the four methods are all flat in the Arctic. Only the PAI (Pairwise Comparison) method indicates adequate temperature change for the sea ice evolution. However, CPS is the only method that has significant ΔT in the Northern Hemisphere temperate latitude band.
The vast majority of alpine/valley glaciers (like the Alps and Glacier National Park) are in the temperate latitudes, formed after the Holocene Climatic Optimum, reaching their peak extents in the mid-1800’s and then generally retreating to their current positions. CPS is the only method that appears to be consistent with this.
If a multiple choice question has five answers:
- Looks correct.
- Possibly looks correct.
- Obviously wrong.
- Obviously wrong.
- Obviously wrong.
Would you pick the answer that looked correct? Or would you average the five answers to come up with a consensus?
Cancelling inconvenient truths
Climate cancel culture has been actively trying to erase the low frequency climate signal ever since the days of Climategate. This is not a new battle:
So, what would it mean, if the reconstructions indicate a larger (Esper et al., 2002; Pollack and Smerdon, 2004; Moberg et al., 2005) or smaller (Jones et al., 1998; Mann et al., 1999) temperature amplitude? We suggest that the former situation, i.e. enhanced variability during pre-industrial times, would result in a redistribution of weight towards the role of natural factors in forcing temperature changes, thereby relatively devaluing the impact of anthropogenic emissions and affecting future predicted scenarios. If that turns out to be the case, agreements such as the Kyoto protocol that intend to reduce emissions of anthropogenic greenhouse gases, would be less effective than thought.Esper et al., 2005
Hockey sticks are usually the result of muting the amplitude of the low frequency climate signal and then splicing on the high resolution instrumental data. The Hockey Team invented cancel culture:
CRU email #1140039406. This email, dated February 15,2006, documented exchanges between several climate scientists, including the Deputy Director of CRU, related to their contributions to chapter six ofthe IPCC AR4. In one such exchange, the Deputy Director of CRU warned his colleagues not to “let [the Co-Chair of AR4 WGl] (or [a researcher at Pennsylvania State University]) push you (us) beyond where we know is right” in terms of stating in the AR4 “conclusions beyond what we can securely justify.”NOAA OIG Report
The CRU’s Keith Briffa was warning his colleagues to not allow NOAA’s Susan Solomon or Penn State’s Michael Mann to coerce them into going along with unsupportable conclusions. This particular e-mail exchange dealt extensively with paleoclimate reconstructions. Briffa also urged his colleagues not to “attack” Anders Moberg, who had recently published a climate reconstruction which actually honored the data and used proper signal processing methods.
Susan Solomon is the NOAA official who claimed that NOAA work related to the IPCC was not subject to FOIA. Michael Mann was the lead author of the thoroughly debunked original Hockey Stick. The late Keith Briffa was the lead author of one of the problematic reconstructions in which “Mike’s Nature Trick” was employed to “hide the decline.” Fortunately scientists like Jan Esper, Anders Moberg, etc. did not succumb to the bullying.
So… Kudos to Kaufman, McKay, Routson and the et al for not deleting the CPS reconstruction to hide the rise and decline of the Holocene Climatic Optimum…
Bohleber, P., Schwikowski, M., Stocker-Waldhuber, M. et al. New glacier evidence for ice-free summits during the life of the Tyrolean Iceman. Sci Rep 10, 20513 (2020). https://doi.org/10.1038/s41598-020-77518-9
Dyke, A.S., Moore, A. and L. Robertson. [computer file]. Deglaciation of North America. Geological Survey of Canada Open File 1547. Ottawa: Natural Resources Canada, 2003.
Esper, J., R.J.S. Wilson, D.C. Frank, A. Moberg, H. Wanner, & J. Luterbacher. 2005. “Climate: past ranges and future changes”. Quaternary Science Reviews 24: 2164-2166.
Grosjean, Martin, Suter, Peter, Trachsel, Mathias & Wanner, Heinz. (2007). “Ice‐borne prehistoric finds in the Swiss Alps reflect Holocene glacier fluctuations”. Journal of Quaternary Science. 22. 203 – 207. 10.1002/jqs.1111.
Kaufman, D., McKay, N., Routson, C. et al. Holocene global mean surface temperature, a multi-method reconstruction approach. Sci Data 7, 201 (2020). https://doi.org/10.1038/s41597-020-0530-7
Stein, R. , Fahl, K. , Schade, I. , Manerung, A. , Wassmuth, S. , Niessen, F. and Nam, S. (2017), Holocene variability in sea ice cover, primary production, and Pacific‐Water inflow and climate change in the Chukchi and East Siberian Seas (Arctic Ocean). J. Quaternary Sci., 32: 362-379. doi:10.1002/jqs.2929 stein2017
via Watts Up With That?
March 26, 2021 at 08:44AM