- JoNova - https://www.joannenova.com.au -

The message from boreholes

There have been suggestions that Jo Nova might be trying to hide or ignore the most recent boreholes graph from Huang et al. So here it is. This is the last 2,000 years according to 6000 boreholes, with the last 100 years also using the “instrumental record” which gives us that hockey-stick uptick at the end. Below I explain the pros and cons of this study and update my thoughts.

Huang and Pollack 2008: Their latest boreholes published study


A borehole sounds like a bit-of-a-stretch as a proxy. How could we tell if the world was warmer in 1066 by drilling a hole in the ground? Yes, fair point. But what makes boreholes useful is that they are global and there is a lot of data: specifically 6,000 holes all over the world.

I’ve been looking at boreholes in more detail, analyzing them in the light of newer proxies. When all the evidence is considered, boreholes turn out be not-much-use at giving us meaningful numbers in degrees C, and in my opinion, not-too-hot at telling us the “when” of an event either. Too much depends on assumptions.

But what are they good for is that, when combined with other proxies, they can help show whether a temperature swing was regional or global.

The basics of Boreholes

Heat from the surface slowly sinks deep into the Earth. Theoretically a hot decade will warm the rocks below and that wave of heat will travel slowly downwards. Rather annoyingly, heat from the boiling hot magma at the center of the Earth is moving up at the same time. As the wave of heat moves down from the top it gradually spreads and blends with cooler decades, information is lost and resolution fades. But if we dig holes down to 2,000m below the surface it’s possible to see signals that appear to be from the Little Ice Age, the Medieval Warm Period, and the Holocene Optimum — going back 20,000 years.

The last ice age seemingly still leaves it’s cold mark on rocks far below, but details like the Younger Dryas spike and the Roman warming are gone, forever blended into the rocks. In other words, there is vastly more detail in ice cores, and we wouldn’t bother with boreholes at all if there were glaciers conveniently located all over the world.

How fast does that heat spread down? In the words of Huang:

Following a change in temperature at the surface, it takes about 100 years for the perturbation to reach a depth of 150m, and 1,000 years to reach 500m depth.

Boreholes are handy because they assess land areas that have few other proxies.

The pro side of boreholes is that there are thousands of measurements, and they are spread all over the land masses of the globe (all bar Antarctica). On the downside, it’s hard to calibrate, and doesn’t include the ocean. After looking closely at these records below, all I can say is that when we have enough data, we’ll be able to use boreholes to tell us highly smoothed past temperatures. Though that might be a thousand years from now.

The boreholes are still useful

For current national policies which aim to change the weather, boreholes suggest three things: 1/ the world has been warmer and colder than now, and sometimes it was global; 2/ those past swings have nothing to do with CO2; 3/ the models can’t explain these swings, so they don’t understand what drives the climate.

Huang and Pollack published a huge study in 1997 with 6,000 boreholes across many continents. What happened over the next ten years in boreholelogy was three different versions of boreholes from the same team all of which contradict each other about the amplitude of the swings.

Let’s follow the curious progression. All these boreholes tell us there were two warm bumps in the last 20,000 years, and two cool swings, but it’s hard to pin down when and how big those bumps and swings were. The basic graph curves seem to flex according to the dominant proxy meme of the day.

Huang et al 1997 “warmer than the present”

Note their conclusions in 1997 — they thought they could compare temperatures to the present day, and to fractions of a degree. They also thought their reconstruction matched the other proxies known at the time.

From the Abstract: “The early to mid-Holocene appears as a relatively long warm interval some 0.2-0.6 K above present-day temperatures, the culmination of the warming that followed the end of the last glaciation. Temperatures were also warmer than present 500-1000 years ago, but then cooled to a minimum some 0.2-0.7C below present about 200 years ago.

Although temperature(s)… are highly smoothed, the results clearly resemble the broad outlines of late Quaternary climate changes suggested by proxies.”

Huang and Pollack have published several newer graphs (posted below) since this one, and now say that this earlier study can’t be used to compare temperatures to modern times, since they had “excluded the data from the top 100m” (which was stated in the ’97 paper), and in this graph below the zer0 point really is around 1900, not 1997. But they have not published a corrected version of this graph.  In 97 Huang said “The uppermost 100m is the depth range most susceptible to non-climatic perturbations such as advective heat transfer associated with ground water flow and terrain effects…”

But it begs the question, if the data from the last 100 years is so useless we need to throw it away, why would older poor resolution records in deeper rocks be that useful? With boreholes, isn’t the best data the data in the most recent waves of heat?

Another point that sets this earlier study apart from later ones is that it was an independent analysis, based only on boreholes and not dependent on any other proxies.

Huang Pollack 1997 boreholes

Last 20,000 years. Huang and Pollack 97. The caption says 1000 years ka b.p. The text refers to “the present” but this graph ends around 1900 according to Huang in 2008.

But lo, the consensus on proxies was changing. In 1998 the infamous Mann Hockey Stick was published, and other proxies were coming out as well. The Mann v Huang graphs didn’t sit well together, to put it mildly.

Tree-rings adapted from Mann et al 1999 ; Boreholes from data from Huang et al 1997 graphed by McKitrick and Essex and adapted here. The modern zero end point was described in 2008 as being around the year 1900. Assuming the other end of the boreholes graph stays at the year 1000AD (though maybe it should line up with 900AD?), I fitted it as best as possible to the centre of the Mann trend at the 1900AD point. Obviously, these two global reconstructions are supposed to be for the same planet. One (or both) was wrong.

Mann was rapidly becoming the star with his bristlecones pines and the hockeystick graph, but what happened to 6,144 global boreholes? Hardly anyone would cite them. The IPCC ignored them, and instead repeatedly promoted the smaller, shorter bristlecone work based on deeply flawed maths and assumptions.

Huang et al 1997 – 2000:  500 years of warming but where did 20,000 years of data go?

The boreholes team of Huang, Shen and Pollack put out quite a few papers from 1997 – 2000, but instead of 6,000 boreholes, these studies dropped 90% of the holes, and 19,500 years of the data. They used between 358 and 616 boreholes. The results were cited widely.

Huang pollack 2000

Note how cold things are in the year 1500?

The main finding that was reported was how steep the recent warming is — but look at the numbers. It’s still only 0.5 K since 1900  — it’s supposed to be 0.9 K according to the thermometers (after adjustments and homogenization).

In the twentieth century alone, the average surface temperature of the continents has increased by about 0.5 K, and the twentieth century has been the warmest century of the past five.

In another paper in the year 2000, the team essentially repeated the 1997 conclusions on page 354 of Ann. Rev. Earth Planet Sci. 2000 28:339-65 so the authors still thought that the world was warmer in 1300 or so. Though oddly in that same paper they also used their heavily truncated 500 year graph, where the year 1500 was the coldest of the cold, which is at odds with the mild temperature of the same year in the 1997 paper. Hmm. Incompatible? Completely.

On a longer timescale embracing all of the Holocene, Huang et al (1997) used the global heat flow database (Pollack et al 1993) to establish a composite profile of heat flow versus depth to 2 km beneath the surface. The inversion of this profile revealed a long mid-Holocene warm interval some 0.2–0.6 K above present day temperatures, and another similar but shorter warm interval 500–1,000 years ago. Temperatures then cooled to a minimum of approximately 0.5 K below present, about 200 years ago.

How could Huang be so specific about the range (0.2–0.6 K) as compared to the present if he and his colleagues were not sure at the time whether the present meant 1900 or 1990?

Then the skeptics asked for the data

McKitrick and Essex asked Huang for his data in 2002 and graphed a close up of that last 1000 years. Since the ’97 paper mentioned “the present” so many times, they naturally assumed that the data ended in 1990. McKitrick and Essex published a very good paper. Their point was that the IPCC cherry picked the study they liked (Mann’s), and ignored the study that fitted the known proxies of the time but did not fit not their pet theory.

Note that Huang in 2008 specifically said that this older graph applies to 1900 not 1990. But it is otherwise still valid. In most proxies (see Loehle 2008 or Ljundqvist 2012), the coldest point of the little ice age was in 1700, and the warmest point of the medieval warm period was somewhere between 800 – 1000AD (with another lesser peak occurred in 1200). Temperatures in 1600 are usually cooler than 1900 in most proxies.  Quite probably this whole graph should be shifted back 100 years — though the text in the paper specifically talks about the peak cooling being “200 years ago”. It’s all untidy. Even if it’s shifted back 100 years, it doesn’t fit the proxies of very warm 9th and 10th centuries. Shouldn’t it blend the two (or multiple) peaks shown in other proxies?

Huang Pollack 1997 boreholes

Last 2,000 years, McKitrick and Essex (book, 2002) Data from Huang. Huang now says that the end point of the graph lines up with 1900 not 1990. There is no corrected graph to post, and the extremes in the paper are described as 200 years ago, and 700-800 years ago. The latest updated 2008 graph suggests the years 800-1000AD was nearly as cold as the Little Ice Age. That doesn’t help either.

Huang 2008 shows “same or slightly less” warmth than the present

In 2008 Huang et al published a new 20,000-year study, but this time used three data sets to get one line. The HP 97 heat flux data set was used again from 20,000 – 300 years ago, but the HP 2000 set was used for 300-100 years ago, and the last 100 years is from the instrumental record from land for the 20th Century. The first graph below covers the last 2000 years. Unlike the 97 study the 2008 one is strongly tied to “the instrumental record”, so it’s not independent and has all the same problems the “instrumental set” has (namely, the micro-siting near tarmac and car parks, urban heat island effects, homogenization, and inexplicable adjustments. See Australian problems, and US ones).  I prefer proxies that continue to “present day”. If the boreholes don’t work for the last 100 years, why would they work for the last 1000?

The graph below and text in the paper suggest the Medieval Warm Period (MWP) was about the same as mid-20th Century temperatures.

The reconstructions show the temperatures of the mid-Holocene warm episode some 1–2 K above the
reference level, the maximum of the MWP at or slightly  below the reference level, the minimum of the LIA about 1 K below the reference level, and end-of-20th century temperatures about 0.5 K above the reference level.

How much warmer is it since the mid-1900’s?

Hadcrut4 would say 0.6C. But the boreholes were done on land, which varies more than the global temperatures, but paradoxically, because the land was warm in the mid 1900’s, that means the land temperatures haven’t changed as much as the global temp since then. Since there are so many problems with the adjusted instrumental record, I no longer think it’s possible to say exactly how much warmer the land temperature is. It’s significant that the mid-20th Century raw thermometer observations from most continental landmasses shows a much smaller rise than the highly adjusted records. The 1930’s and 1940’s were hot on land in most continents. Not just in the US.

Note that Huang and Pollack have a suite of reconstructions to pick from (below), but while most proxies are suggesting there were two peaks in the MWP (the biggest before 1000AD and the second around 1200AD) the  borehole reconstruction below that suggests things were “as warm” in the year 1000AD (q60a09) is also the one that tells us it peaked in 1200. The other reconstructions are too early, and unrealistically cold in the year 1000AD. They are also much colder in the Little Ice Age than other proxies suggest.

Huang Pollack 1997 boreholes

Huang et al 2008. Note the peak of the MWP is around 1300, which is not the same as most land proxies suggest. The latest multiproxies suggest the peak was before the year 1000AD. The curves closer to that are also the highest. The zero K horizontal line is the reference period 1961-1990.

Is Pollack an un-skeptical scientist?

Henry Pollack published a book “Uncertain Science, Uncertain World”  in July 2005. In official reviews, his book appealed to the IPCC head  Sir John Houghton, and another well known alarmist Stephen Schneider of Stanford University: “This well-written book is a welcome antidote to the misrepresentations of special interests who misuse scientific uncertainty to stall public policy and advance their own agendas.”

Pollack’s points about uncertainty may well be valid, but the point of the book was to “demolish the mythology” that uncertainty is an excuse for inaction. Pollack was not sympathetic to skeptical scientists, and it’s fair to ask whether he was as dispassionate about the results as a scientist should be. Would anyone who felt “climate action” was necessary be pleased to report results that showed climate action is a fruitless endeavor?

Oh look, in 2010 he wrote “A world Without Ice” and the book description tells us the author is A cowinner of the 2007 Nobel Peace Prize (and we all know how accurate that is). “The arctic is imperiled”.

The bottom line

Huang 2008, for what it’s worth, shows that the world was warmer in the Holocene 5,000 years ago, even warmer than our current airports and carparks  are today; warmer than the air next to air-conditioner exhaust vents. Ice cores show CO2 levels were lower. The central point keeps coming back: the models don’t know what caused that warming, it wasn’t CO2, and we know it was not a disaster for life on Earth either.

Huang Pollack 1997 boreholes



 The borehole’s monster spaghetti graph

Here are all three graphs, none of which agree with each other, or with the hockey stick or any other major proxy. Most proxies put the depth of the little ice age around 1700, and the peaks of the MWP either 900- 1000 or 1200 years ago. Nothing fits.



It possibly says more about the political nature of climate research than anything much about our climate.

Last word

What do boreholes tell us? That past warming episodes were global. That the IPCC cherry picks studies to suit it’s political mission.

In order to see if medieval temperatures were warmer cooler or the same, we need to look at other proxies, which I will do soon.



Huang, S., H. N. Pollack, and P. Y. Shen (1997), Late Quaternary temperature changes seen in world‐wide continental heat flow measurements, Geophys. Res. Lett., 24(15), 1947–1950. [Abstract, PDF] Discussion

Huang, S., H. N. Pollack, and P. Y. Shen (2000), Temperature trends over the past five centuries reconstructed from borehole temperatures, Nature, 403, 756– 758. [PDF]

Huang, S. (2004), Merging information from different resources for new insights intoclimate change in the past and future, Geophys. Res. Lett., 31, L13205, doi:10.1029/2004GL019781.

Huang, S. P., H. N. Pollack, and P.-Y. Shen (2008), A late Quaternary climate reconstruction based on borehole heat
flux data, borehole temperature data, and the instrumental record, Geophys. Res. Lett., 35, L13703, doi:10.1029/2008GL034187   [PDF]


 Stoat has a useful thread on this from way back in 2007 with some good comments.

Hat tip to NiceOne who drew my attention to the contradictions in the Boreholes studies and details of the updated version.

Start the countdown to the argument that only “experts” can understand data this complex. 3.2.1….

8.3 out of 10 based on 65 ratings