Documenting the Global Extent of the Medieval Warm Period

Repost from Watts up with That here: WUWT

Some typos in the text and diagrams have been corrected

Introduction

In this article I pose the following questions:

  • Was the Medieval Warm Period (MWP) a global event?
  • Were the MWP temperatures higher than recent times?

The reasons for asking these questions are that climate establishment have tried to sideline the MWP as a purely local North Atlantic event. They also frequently state that current temperatures are the highest ever.

I attempt to answer these questions below.

Mapping Project for the Medieval Warm Period

I use the mapping project for the Medieval Warm Period (MWP) developed by Dr Sebastian Luening and Fritz Vahrenholt to establish the global extent of the MWP. This project is a considerable undertaking and I commend the authors for their work.

Luening states on Research Gate that,

“The project aims to identify, capture and interpret all published studies about the Medieval Climate Anomaly (Medieval Warm Period) on a worldwide scale. The data is visualized through a freely accessible online map: http://t1p.de/mwp.”

I show a screenshot from the project in Figure 1 but I recommend that you use the online version at http://t1p.de/mwp because it contains a wealth of data, including abstracts and links to the individual papers.

(Note – The t1p.de/mwp URL given may or may not be reachable by some viewers as it’s one of those mangled “URL shorteners” that may or may not work. It does not work for me. The author claims the URL is valid, and offered no alternate, I independently found this link to a Google maps document which might work better for some readers – Anthony)

fig-1-screenshot-of-mwp-project

key-to-figure-1

Figure 1: Screenshot of MWP Mapping Project (Source: Luening http://t1p.de/mwp downloaded 27-Dec-2016)

A cursory inspection of Figure 1 indicates that there are a large number of warm study locations dispersed throughout the world. However, to determine the global numbers for the Warm-Cold-Neutral-Dry-Wet studies, I downloaded the mapped data for the 934 studies that were available on 30 December 2016 and these are summarised in Figure 2.

fig-2a-temperature-hydroclimate-data

 

fig-2b-temperature-data-only

Figure 2: Results from MWP Mapping Project (Source: Luening http://t1p.de/mwp downloaded 30-Dec-2016)

The following observations are evident from Figure 2;

a. The number of Warm studies (497) greatly exceed the other studies, namely, 53% of the studies when temperature and hydroclimate data is used and 88% when temperature only data is used.

b. The number of Cold studies (18) is very small, at 2-3% of the overall studies.

c. The number of Neutral studies (53) is comparatively low, at 6-9% of the overall studies.

d. The number of studies that report only Hydroclimatic data is not insignificant. The number of Dry studies (184) and the number of Wet studies (182) are 20% and 19% of the overall studies respectively.

In summary, the overwhelming evidence from the Luening MWP Mapping Project to date is that the MWP was globally warm but it is not immediately obvious what the definition of warm is?

Descriptions of warm or cold are given the individual studies. For example, Kuhnert & Mulitza (2011: GeoB 9501) (extracted from Luening) states that the,

“Medieval Warm Period 800-1200 AD was about 1.1°C warmer (50 years mean) than subsequent Little Ice Age.”

Now, whilst this description is useful, it does not allow us to compare MWP temperatures with modern warming. Therefore, I compare modern temperatures with the MWP below.

How Warm was the MWP?

Several of the references in the Luening MWP Mapping project are also so referenced in Ljungqvist (2010), therefore the latter may be used to estimate the temperature in the Northern Hemisphere (NH) during the MWP.

Note that I have used the Ljungqvist (2010) NH paper because I already had the data available on my laptop. However, readers could carry out a similar exercise by using their preferred study to derive appropriate temperatures for global or hemispherical temperatures for comparison with the present day.

Ljungqvist published his data along with his paper (other climate scientists should do the same) and Figure 3 shows my version of the Ljungqvist chart (reproduced using his data). I have also annotated the chart to show previous warm and cold periods.

fig-3-estimate-of-nh-temperature-ljungqvist-2010

Figure 3: Estimate of Extra-tropical Northern Hemisphere (90–30°N) Decadal Mean Temperature Variations (after Ljungqvist, 2010)

a. All values are decadal mean values.

b. Proxy temperatures are shown in the thick black line and are relative to the 1961–1990 mean instrumental temperature from the variance adjusted CRUTEM3+HadSST2 90–30°N record.

c. Two standard deviation error bars are shown in light grey shading.

d. CRUTEM3+HadSST2 instrumental temperatures are shown in the red line.

The following observations are evident from Figure 3:

a. The Medieval Warm Period (MWP) temperature is warmer than the Modern Warm Period proxy temperature – at no time do the modern proxies exceed the MWP proxies.

b. The recent instrumental temperature is warmer than the Medieval Warm Period temperature.

c. The MWP is obviously preindustrial and it would be logical to use the MWP temperature as preindustrial temperature instead of the Little Ice Age (LIA) favoured by the climate establishment.

Observations (a) and (b) above are based on Ljungqvist’s comments. He states that [emphasis added],

“…the Medieval Warm Period, from the ninth to the thirteenth centuries, seem to have equalled or exceeded the AD 1961-1990 mean temperature level in the extra-tropical Northern Hemisphere.”

He also adds the following words of caution [emphasis added],

Since AD 1990, though, average temperatures in the extra-tropical Northern Hemisphere exceed those of any other warm decades the last two millennia, even the peak of the Medieval Warm Period, if we look at the instrumental temperature data spliced to the proxy reconstruction. However, this sharp rise in temperature compared to the magnitude of warmth in previous warm periods should be cautiously interpreted since it is not visible in the proxy reconstruction itself.”

Regarding Ljungqvist’s note of caution and item (b) above, comparing instrumental temperatures with those derived from proxies is not an, “apples for apples” comparison because proxy temperatures are damped (flattened out) whilst thermometers respond very quickly to changes in temperature.

Ljungqvist describes the dampening proxy temperatures thus [emphasis added],

“The dating uncertainty of proxy records very likely results in “flattening out” the values from the same climate event over several hundred years and thus in fact acts as a lowpass filter that makes us unable to capture the true magnitude of the cold and warm periods…What we then actually get is an average of the temperature over one or two centuries.”

However, the flattening out of proxy records is not the only problem when comparing proxies with instrumental temperatures – there is also the “divergence problem.” We examine this for the Ljungqvist’s 1850-1989 calibration period in Figure 4.

fig-4-comparison-of-proxy-instrumental-temps

 

Figure 4: Comparison of Proxy & Instrumental Decadal Mean Temperature Variations (after Ljungqvist, 2010)

The decadal correlation (r = 0.95, r² = 0.90) between proxy and instrumental temperature is very high during the calibration period AD 1850-1989. However, note that the 1990-1999 instrumental peak is not used in the calibration. Probably because Figure 4 shows that the instrumental peak temperature is 0.39°C for 1990-1999, which diverges from the proxy temperature of 0.06°C – a difference of 0.33°C.

This divergence is outside the 2 standard deviation error bars of ±0.12°C reported by Ljungqvist and it illustrates the divergence problem in a nutshell – proxies do not follow the instrumental record for recent high temperatures. The reason for this divergence is that the proxy response is probably nonlinear instead of the linear response that is assumed in all climate reconstructions to date.

Ljungqvist explains the nonlinearity [emphasis added],

“One circumstance that possibly has led to an underestimation of the true variability is that we must presuppose a linear response between temperature and proxy. If this response is nonlinear in nature, which is often likely the case, our interpretations necessarily become flawed. This is something that may result in an underestimation of the amplitude of the variability that falls outside therange of temperatures in the calibration period.”

As a structural engineer, I agree with Ljungqvist that nonlinearity is often likely to be the case for proxies. Most natural materials, e.g., timber, rock, etc., exhibit linear-nonlinear behaviour. However, this concept is difficult to explain in words, therefore, I explain the concept diagrammatically in Figure 5.

fig-5-indicative-nonlinear-response

 

Figure 5: Indicative Nonlinear Response for a Temperature Proxy

For clarity, I show the positive part of the temperature response in Figure 5 but it should be noted that there would be a similar negative response. I now explain Figure 5 as follows:

a. Line OAB represents the linear proxy response to temperature currently used in reconstructions. Parameter P1 yields a temperature T1 (via point A) and the response increases linearly so that parameter P2 indicates a higher temperature T2 (via point B).

b. Line OACD represents the indicative nonlinear proxy response to temperature. Parameter P1 is in the linear portion of the curve and therefore results in the same temperature T1 as that in item (a) above.

c. In addition to item (b), parameter P1 at point C is in the nonlinear portion of the curve, which would result in the temperature T2, which is significantly higher than the value of T1 obtained from the linear response via point A.

In summary, a nonlinear proxy-temperature response can result in two temperature values for a single proxy value, which could be problematic in deciding which temperature result to use.

However, for modern temperatures, the nonlinear response of the proxy can be calibrated against the instrumental record. For example, if the temperature results in Figure 4 were for a single proxy then the following procedure could be used:

a. The nonlinear curve would be calibrated so that temperature T1 = 0.08°C would be in the linear portion of the curve in Figure 5.

b. The temperature T2 = 0.39 °C would be calibrated to be in the nonlinear portion of the curve in Figure 5.

Note that the linear response used in Figure 4 shows T2 = 0.06°C but the nonlinear response procedure above allows the calibration T2 = 0.39°C. This type of nonlinear calibration would mean that historical temperatures would be higher than those currently estimated by using a linear proxy-temperature response.

Furthermore, the calibration in modern times is relatively straightforward because we can compare proxies with instrumental temperatures to determine which temperature is correct, T1 or T2. However, historical temperatures are more problematic because we do not have instrumental records for several thousands of years to determine the correct temperature, T1 or T2. Nevertheless, we can estimate the correct temperature by deploying a similar methodology to that used by earthquake engineers as explained below.

Earthquake engineers routinely design structures for 1-in-500 year seismic events and for special structure (e.g., nuclear installations) 1-in-5,000 year events. However, they have only had the seismograph to measure the strength of earthquakes from 1935, when it was invented by Charles Richter. Earthquake engineers solve this short instrumental record dilemma by investigating historical descriptions of seismic events.

For example, the historical record for an earthquake may include a description such as, “The earth boiled.” This literal description does not mean that the earth actually reached boiling. It means that the groundwater in the soil bubbled to the surface – a process known as liquefaction of the soil. Earthquake engineers can then use this description to estimate the amount of shaking that would cause liquefaction and thus estimate the strength of the earthquake at that location.

In climate science, the temperature, T1 or T2, could also be estimated from historical descriptions by using changes in vegetation, human habitation, etc. For example, if our proxy were in Greenland then the following outcomes are possible:

a. If the land were in permafrost at the date of the proxy then the lower temperature T1 would be used.

b. It the land was being cultivated by Vikings at the date of the proxy (i.e., no permafrost) then the higher temperature T2 would be used.

The above nonlinear methodology would obviously require more work by paleo scientists when compared with the current linear calibration methods but this work is no more than is routinely carried out by earthquake engineers.

Additionally, up-to-date proxies for the post-2000 period would be useful to determine if the divergence between proxies and instrumental temperatures has increased. The proxies would also enable a more accurate nonlinear proxy-temperature calibration, say, point D in Figure 5. Strangely, there seems to be a reluctance in the climate establishment to publish up to date proxies.

Using a nonlinear proxy-temperature response would result in historical high temperatures being calculated to be higher and historical low temperatures being calculated to be lower.

Conclusions

A review of the global extent of the MWP is presented and the following conclusions are offered:

  1. The MWP was a global event and a large number of studies show that warming events overwhelming outnumber cold events.
  2. However, the not insignificant number of dry or wet events recorded in the MWP Mapping Project would suggest that perhaps the Medieval Climate Anomaly would be a better description than the MWP.
  3. NH temperatures during the MWP were at least as warm those in 1980-1989 instrumental record.
  4. Recent instrumental temperatures show higher temperatures when compared with the MWP proxies. However, instrumental temperatures should not be compared directly with proxy temperatures because this is not an “apples for apples” comparison. Proxy temperatures are dampened (flattened) out on decadal or greater scales.
  5. Recent proxy records diverge from instrumental temperatures – instruments show higher readings when compared with proxies.
  6. The divergence problem in item (5) above is probably due to a linear proxy-temperature response being assumed in current temperature reconstructions. A nonlinear proxy-temperature response would achieve more accurate results for historical high and low temperatures and achieve a better correlation with recent instrumental data.

Until there is a good correlation between instrumental temperatures and proxies, no reputable scientist can definitely state that current temperatures are the highest ever.

Afterword

Regarding the 2°C limit proposed by the climate establishment, the following phrase (based on the facts presented in this article) does not sound nearly as alarming as is currently promulgated in the media:

“We know that the preindustrial MPW temperatures exceeded those of the LIA by ≈ 1°C without CO2 but we also think that a similar temperature rise from the LIA to 1980-1989 was caused by CO2. Therefore, we think that we need to limit CO2 emissions to stop a further (dangerous) 1°C rise. We think.”

I think that there is too much thinking going on and not enough real climate science.

A good start would be to use nonlinear proxy-temperature responses to establish accurate historical high and low temperatures.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s