The trouble with the surface record (aside from imperfections in data collection including special difficulties with the collection of data over the oceans, and the inadvertent inclusion of non-greenhouse anthropogenic impacts such as land-use changes and heat-island effects) is the lack of a complete global coverage; this is particularly so poleward of 50 S and 70 N.

Gray (2000) explains that the method of arriving at global temperature change involves dividing the Earth’s surface into 5 x 5 degree  latitude/longitude boxes, which are separately evaluated according to the data available before their compilation into a wider record.  He reports that for the period from 1901, only 46% of the boxes contain useable data, and 60% of these are in the Northern Hemisphere.  Worse, the more-reliable land data covers only 14% of the surface over that period - 11% in the Northern and 4% in the Southern Hemisphere.  Furthermore, he suspects that this limited land data has been subject to unrecognised human influences such as urbanisation.

However, for the purposes of this paper, the surface record is accepted at face value: ie it is here treated as an adequate representation of global near-surface temperature variation over the past century and more.

By inspection of Figure 1(c), the global record is seen to comprise two periods of warming (each about 0.5 0C) during 1909-44 and 1977 to the present.  These two prominent tranches of warming are preceded by cooling between 1878 and 1908, and separated by further cooling at 1945-76.  Over the entire 140 years, the rise in global surface temperature is some 0.6 +/- 0.2  degrees C.

However, depending on how the data is treated, different authorities come to slightly different results.  Jones et al (1999) describe the global warming as two tranches of 0.37 degrees C in 1925-44, and 0.32 degrees C in 1978-97, with warming continuing thereafter.  The National Research Council of the United States (Panel on Reconciling Temperature Observations, 2000) puts it a little differently; it sees a warming of 0.053 to 0.059 degrees C/decade for the period 1890-1998, including an accelerated warming of 0.13 to 0.19 degrees C/decade for the 20-years 1979-98.

However, all sources find that the 20th century has warmed, much as shown in Figure 1.

  3.1.2  A natural oscillation in surface temperatures
Schlesinger and Ramankutty (1994) recognise oscillations in surface temperature on a 50-to-88-year period for 11 geographical regions of the world, using the then-available instrumented record from the 1850s to 1992.  They derive an overall statistical result of 65-70 years periodicity. This cycle is easiest to see in the Northern Hemisphere - in Figure 1(a) - over-printed on the longer-term warming trend.

Kerr (2000) has revisited this topic, but without referring to the earlier findings.  He reports a roughly 60-year climate oscillation (which he suggests be called the Atlantic Multidecadal Oscillation or AMO) in and around the North Atlantic within a “total global warming from 1860 to the present of about 0.6 degrees C”.

What is the cause of this oscillation?  Schlesinger and Ramankutty suggest that:
 ..... the oscillation arises from predictable internal variability of the ocean- atmosphere system.

Kerr canvasses several possible explanations, for instance:
 Some researchers, particularly climate modellers, suspect that oscillations in the  heat-carrying currents of the North Atlantic are to blame for this natural mode.

I am with the modellers.

An oscillation with such a long period and such a pronounced amplitude is likely to be ocean-related (although neither solar nor lunar impacts can be ignored).  The atmosphere cannot ‘remember’ unaided from one year to the next, let alone for 60-70 years.  As discussed later, the reason that the influence on climate is most obvious in the Northern Hemisphere relates to basinal geometry in the Atlantic.

 3.2  Explaining the instrumented record: 1860-1944
During this period, the climate records of the hemispheres are markedly different, with that for the Southern Hemisphere displaying much the lesser contrast.  The inference is that more than one factor was driving climate change during this period, at least one of which was not of global coverage.  What are these factors?

  3.2.1  A role for the Sun?
A logical start is to consider the Sun.  Insolation during the 11-year solar cycle typically varies by only about 0.1% on a peak-to-trough basis; and even during the infamous Maunder sunspot minimum of 1645-1715, solar output was only about 0.4% less than now.  These are small fluctuations.

However, Parker (1999) provides a plausible explanation for the prominent, and long-recognised, correlation between solar activity and climate, saying:

Finally, it has come to light that cloud cover follows the cosmic-ray intensity  closer than (it does) any direct index of solar activity.  We know that the  condensation of water vapour to form ice crystals and water drops is nucleated  by ions in the atmosphere, and cosmic rays are the principal source of  atmospheric ions.  The intensity of cosmic rays is strongly suppressed by solar  activity, so any increase in activity means reduced nucleation opportunities for  cloud formation and an associated increase in sunlight reaching the surface of  the Earth.

Fluctuation in cosmogenic 10Be and 14C abundance is directly attributable to variations in the interplanetary magnetic field - the stronger the field, the less cosmic rays penetrate to Earth, and the less is the consequent production of radio-isotopes in the earth’s atmosphere.  These isotopes are present in ice-cores and tree-rings, and it is therefore possible to reconstruct an intensity record for the solar magnetic field.

It is now understood that the weaker the Sun’s magnetic field, the greater is the extent of cover by low clouds (below 3-4 km) in the Earth’s atmosphere.  In climate-change terms, this is a positive feedback in response to variable solar activity.  Herein lies the explanation of the observed colder climate at times of low sunspot numbers: it is the result of increased cloudiness.

Lockwood et al (1999) note that the total magnetic flux leaving the Sun has risen by a factor of 1.4 in the period 1964-96, and (here using surrogate measurements) by a factor of 2.3 since 1901.  Figure 2 (a) shows a more-recent and longer reconstruction of solar magnetic flux extending back to 1700 and the end of the Maunder minimum.

Figure 2(b) compares a somewhat earlier reconstruction of solar activity with Northern Hemisphere on-land surface temperatures back to 1750 (such long sea-surface and Southern Hemisphere records are not available).  This correlation, and a comparison of Figure 2(a) with Figure 1(c), indicate that the Sun has played a major role in climate change both before and since 1860.

3.2.2  A role for the oceans?
But there is more.  Figure 3 contains three portrayals of the last cycle (Mediaeval Warm Period AD800-1300, and Little Ice Age 1300-1900) in the ca 1500-year warm/cold pacing of North Atlantic Basin climate which has persisted right through the 10,000 years (10 ky) of the present (Holocene) Interglacial, and for many tens of millennia in the Glacial which preceded it.

The left-hand graph in Figure 3 (see Dahl-Jensen et al 1998) shows the measured palaeotemperature in a core-hole on the Greenland ice-sheet at 65 0N; the centre graph is a proxy record from an Atlantic sea-bed core at 30 0N (Keigwin & Pickart 1999); and a similar record from 20 0N (deMenocal et al 2000) is on the right.  All sites display a twin-troughed Little Ice Age.
 
Figure 4 displays a proxy for the flow of warm Atlantic water into the Nordic seas during the last 10 ky of benign Holocene climate.  The Mediaeval Warm Period and the subsequent Little Ice Age can be distinguished.  As might be expected the LIA, at about AD1300, begins with a reduced northward flow of warm water.  (In fact, the graph is based on a proxy for the variable returning flow of cold North Atlantic Deep Water across the Iceland/Scotland sill into the Atlantic.)

Figure 5 presents the other face of the same coin.  Increasing occurrence of the planktonic foraminifer Globigerina bulloides appears to reflect the upwelling of colder, nutrient-rich, waters to the surface.  Here, a core in the Cariaco Basin (at 11 degrees N 65 degrees W, in the southern Caribbean) plots the abundance of G. bulloides over the past 850 years.

While it is difficult to separate out all the factors influencing this fluctuating record (including variations in solar activity), one feature stands out.  Richness of the marker foram in the cored sediments begins an abrupt decline about AD1320, and does not start to recover until after 1380.  This interval corresponds both to the beginning of cold conditions in the North Atlantic (Figure 3) and to the reduction of warm water flow into high northern latitudes (Figure 4).  But reduced abundance of G bulloides is more likely to indicate warming than cooling in the Caribbean at that time.

This countercyclic warming of the Caribbean tends to confirm that the event triggering the start of the Little Ice Age at about AD1300 is not primarily a change in solar activity.  It appears much more likely to be related to a contemporaneous change in oceanic heat transportation.  Increased diversion of warm equatorial water to the Caribbean has suppressed upwelling therein, and has concurrently starved higher latitudes of warmth.

Climate cyclicity during the Holocene, and the similar but much more prominent fluctuations during the previous Glacial, are associated with the outburst of iceberg armadas into the ocean.  Continental ice at high latitudes translates, once afloat, into sea-level rise around the world including equatorial regions.  The Globe’s radius of gyration increases as a consequence and, in order to preserve its level of angular momentum, our planet must spin more slowly.  The concurrent warming of the Caribbean and cooling of the North Atlantic is an inertial effect (of which more later).

  3.2.3..A role for the greenhouse effect?
What of greenhouse warming in the first half of the 20th century?  Carbon dioxide is the dominant anthropogenic greenhouse gas; and ice-core data indicates that its pre-industrial (early 1800s) concentration in the atmosphere was 280 ppm, rising to 308 ppm by 1944 at the end of the earlier tranche of warming.  It was still only 315 ppm when continuous monitoring began in 1958, and it is about 368 ppm now.

Two-thirds of the increase in atmospheric CO2 concentration since the early 1800s post-dates 1944.  Either global temperature rose in anticipation of the CO2 emissions, or we must look elsewhere for the explanation of most of the 1860-1944 warming.
  3.2.4  First tranche of 20th century warming: the drivers
Natural variability - particularly rebound from the inertially-related Little Ice Age (in the form of a reinstated stronger northerly flow of warm water in the North Atlantic) and increasing solar activity (in the form of a positive feedback from reduced cloudiness) - provides a much more-plausible explanation of observed warming in the first half of the 20th century (and earlier) than does human-caused changes to the composition of the atmosphere.

 3.3 Explaining the instrumented record: 1945-2000
  3.3.1  The surface record
At the surface, the later record divides readily into a period of gradual cooling from 1945 to 1976, and a period of renewed warming thereafter (Figure 1).  But warming is far from uniform in its distribution.  Figure 6 (reproduced in colour, and bound at the front of this paper) depicts global surface temperature changes during the 1969-98 period.

The greater warming is in the Northern Hemisphere, and (at least for the past two decades) two-thirds of this northern warming takes place in winter.  More remarkable though is the distribution of the winter warming.  It is mostly in the intensely cold and very dry high pressure systems of Siberia and Alaska/Yukon, where temperatures remain far below freezing even after the warming.  In these cold/dry regions, winter half-year warming is 0.21 degrees C/decade cf an average of only 0.02 degrees C/decade for the greater part of the Hemisphere.

  3.3.2  Balloon-borne thermometers
An important development in the latter part of the 20th century is the availability of temperature data for the lower atmosphere.

Radiosondes (weather balloons) provide a continuous record of temperatures in the lower atmosphere back to 1958.  However, their coverage is almost entirely restricted to land areas, and also excludes the polar regions.

There is a mystery apparent in the balloon record, as shown in Figure 7.  There is almost no warming trend in atmospheric temperatures prior to 1976/77, nor thereafter; and virtually all the substantial jump of about 0.3 degrees C occurs in a single year!

The rate of greenhouse warming predicted by climate models tends to be fairly constant.  Hence whatever its cause, this step-change is unlikely to be the result of human interference with the composition of the atmosphere.  This is a crucial issue in the context of my submission, and I return to it later.

However, despite its incomplete and selective coverage, the balloon record appears to acknowledge the known major short-term influences on global climate.
In particular, it appears responsive to the El Niño/Southern Oscillation (ENSO) fluctuations (see later), centred in the equatorial Eastern Pacific.  In addition, this record shows a marked cooling in the years immediately following the 1963 Mt Agung and 1991 Mt Pinatubo eruptions.  This circumstantial evidence suggests that the balloons are providing valid information on global atmospheric temperatures.  (Interestingly, the 1982 El Chichón eruption and the large 1983 El Niño warming event are not prominent.  They appear to have, in part, cancelled each other.)

3.3.3  Satellite-borne microwave sounding units
A world-wide coverage of temperature change has only been available since January 1979.  This record (also for the lower atmosphere) is derived from microwave sounding units borne by polar-orbiting weather satellites, and the global monthly average is shown in the top graph of  Figure 8.  On a global basis, there is good agreement between lower atmosphere temperatures obtained by radiosondes and satellites during their period of overlap.

All or most of the fluctuation in globally-averaged temperature for the lower atmosphere over the past 22 years (the duration of the MSU record) can be explained in terms of volcanic cooling and ENSO-related variations in Pacific sea-surface temperatures.  Over this period, there is little apparent global trend in the temperature of the lower atmosphere (average warming of about 0.05 degrees C per decade) even when the major El Niño event of 1997/8 is included.

3.3.4  North/south differences in the atmosphere
The lower pair of graphs in Figure 8 separate out the Northern and Southern Hemisphere records; clearly the two hemispheres are subject to the same drivers.  However, there are differences.  For instance, the impact on atmospheric temperature of the 1980 El Niño appears to have been suppressed in the north, but not the south, by the Mt St Helens eruption (5/80, 46 degrees N).  Even cooling events originating much nearer the equator, such as that at El Chichón in Mexico (3/82) or Mt Pinatubo on Luzon (6/91) at 15-20 degrees N, are more strongly expressed in the Northern Hemisphere.

Despite an obviously-lesser volcanic cooling influence in the Southern Hemisphere, it still displays a slight cooling trend in the lower atmosphere (0.04 degrees C) over the duration of the 22-year record, which partly cancels the warming trend observed in the Northern Hemisphere (0.13 degrees C).

The unexpected direction of the divergence in this inter-hemispheric trend has important implications, because IPCC’s purported validation against the past of the models used to project future climate change, invokes cooling by anthropogenic aerosols (which are mostly emitted in the Northern Hemisphere).  IPCC needs cooling aerosols to bring its over-predicting models into line with an under-warming world.

3.3.5  Atmosphere-surface comparisons
As can be seen in Figure 8, there is much fluctuation but little trend in the temperature of the lower atmosphere.

But Figures 9 and 10 show that, although both surface and atmosphere are responding to the same influences, there is an underlying warming trend at the surface which is hardly present in the atmospheric record.

This remarkable dichotomy is crucial to any plausible hypothesis as to what is driving climate-change.

You read it first here

© 2001  Bob Foster  Posted   9, April, 2001
www.globalwarming-news.com
Back to "Duel of the Hypotheses"  contents page
Back to Guests Page
Back to Front Page