On the recovery from the Little Ice Age

by Syun-Ichi Akasofu

1. Introduction: The Little Ice Age (LIA)

Significant data show

  • the LIA certainly ended in about 1800-1850 and
  • the recovery has continuously progressed to the present with superposed ‘fluctuations’

Let’s first review changes of temperature from about 1000 to the present.

Figure 1(a) shows a typical example of tree-ring data from the middle latitudes. Compared with the mean 1961-1990 level, the temperature was relatively low from about 1100 to 1800-1850, indicating that the Earth experienced a relatively cool period, the LIA.

The recovery from the LIA began in about 1800-1850 and temperature has increased from then to the present. The temperature rose from 1800-1850 to the present continuously with superposed ‘fluctuations’ and show no sign the recovery ended before 1900.

Figure 1 (c) shows the ice breakup at Lake Suwa in the central highland of Japan from 1450 to 2000.

The lake has a nearly circular shape, and this particular break-up phenomenon, called “Omiwatari,” meaning ‘God’s crossing’, tends to occur during the early freezing period, perhaps because of the pressure exerted by the expanding ice. The breakup produced a loud sound causing people to think God crossed the lake. For this religious reason they kept a long record of this breakup. The delay of the break-up dates indicates warming from 1800 to the present. This shows the LIA occurred in Asia and many publications indicate that the LIA was a worldwide phenomenon.

2. The Recovery from the LIA

In this section, we learn about climate change from 1800-1850 to 1900. Since we have much more accurate data after 1900, climate change after 1900 will be dealt with in the next section.

We find ice core data, river freeze/break-up dates, sea level changes, sea ice changes, glacier changes, tree-ring data and cosmic-ray intensity data show the recovery from the LIA progressed regularly from 1800-1850 to the present.

2.1. Ice Core Data

Figure 2(a) shows a continuous rise of temperature from 1775 to 2000, as deduced from ice cores at Severnaya Zemlya, an island in the Arctic Ocean. This record is particularly valuable because we do not expect any contamination by human activities.

The middle curve shows thermometer temperature records from Vardo in northern Norway. the bottom curve shows temperature changes at stations along the coastline of the Arctic Ocean. These thermometer records support the credibility of the ice core record. Polar amplification of the multi-decadal oscillation caused the large positive change from 1910 to 1975, which will be discussed in Section 5.

2.2. River Freeze/Break-up Dates

Data of a number of lakes and rivers of the world from 1846 to 1995 show break-up dates have almost steadily advanced to earlier dates while freeze dates shift to later dates.

There are also various reports about advancing glaciers during the LIA in Scandinavia. Therefore, it is clear that many glaciers advanced during the LIA before starting to retreat in about 1800-1850. Altogether, long-term glacier data presented show that glaciers advanced from about 1400 and began to retreat rather steadily after 1800-1850.

These facts confirm that the Earth experienced the LIA and began to recover from it as evidenced by a number of natural phenomena, such as retreating glaciers and sea ice from about 1800-1850 to the present.

2.3. Sea Level Changes

The linear trend of the recovery from the LIA can also be seen in sea level changes.

Figure 2. (c) Global sea level change from 1700 to the present. It is clear that the sea level began to increase in about 1850 and continued rising almost linearly to the present.

2.4. Sea Ice Changes

There is no accurate Arctic Ocean data until satellite observations became available in the 1970s. The only long-term observation of sea ice is available from the Norwegian Sea. The large ‘fluctuations’ between 1910 and 1975 are likely related to the multi-decadal oscillation, which is discussed in Section 5.

Figure 2(e) shows variations of the occurrence of sea ice on the coasts of Iceland the year 800. The decline after 1800 corresponds to the northward shift shown of sea ice. The gradual build-up of sea ice, beginning in about 1200 or after 1400, indicates the beginning of the LIA.

2.5. Glaciers

Figures 3(a-f) show records of glaciers in Alaska, New Zealand, the European Alps, and the Himalayas. These glaciers have been receding from the time of the earliest available records, as shown by accurate terminus records. There are also a large number of similar records from the European Alps, Alaska, and elsewhere. Thus, many glaciers in the world have been retreating from 1800-1850 to the present.

Therefore, glacier retreat is not connected to human influences.

A large number of recent publications show photographs of the same glaciers taken early and late in the 1900s. AGW supporters have claimed these photographs are evidence of the effect of CO2. However, Figure 3(a-f) demonstrate that those photographs are misleading and are not evidence of the sudden warming after 1900 and of the greenhouse effect. Therefore, such photographs cannot be used as evidence to support the greenhouse effect of CO2.

Let’s examine glacier changes before 1800.

Figure 3. (a) Retreat of glaciers in Glacier Bay, Alaska, began in 1800.

Figure 3 (b) Retreat of the Franz Josef Glacier in New Zealand began before 1865.

Figure 3 (c) Retreat of the Gangotri Glacier in the Himalayas began before 1800.

Figure 3(d) shows radiocarbon datings of glacial advances in the Juneau outlet. Each advance killed trees and left stumps for analysis. These advances occurred before Glacier Bay glaciers began to recede in about 1800.

Figure 3(f) shows changes of the Mer de Glace glacier in the Alps, which began to build up after 1550 (during the LIA) and began to retreat about 1852.

Historical documents are also available that describe cool weather conditions during the LIA, such as freezing of the River Thames in the 1600.

3. Continuation of the Recovery

In Section 3, having more accurate data after 1900, we learn that temperature changes during the 20th century can be judged as a continuation of the recovery, approximated by a linear change at the rate of about 0.5 C/100 years, with the superposed multi-decadal oscillation.

3.1. The Linearity of the Recovery

In the previous section, we learned that the recovery from the LIA was continuous, although there are super-posed ‘fluctuations’, which will be discussed in Section 5. In this section, we learn, on the basis of more accurate data gathered during the last century, that climate change examined in the previous section has continued to the present. With these data, we can examine more carefully the changes and, specifically the linearity of the changes.

A recent study of sea level changes is shown in Figure 4(a). It shows the last part of Figure 2(c).

Figure 4 (a) shows the mean sea level record from nine tide gauges for 1904-2003 based on the decadal trend values for 1907-1999. The rate of sea level rise has been about 1.7 mm/year and reflects the thermal expansion of seawater and glacier melting during the last half century.

Comparing Figure 4(a) and Figure 2(c), we see that the recovery from the LIA is a continuous process, without major change of the rising rate. This shows the linearity of the change from 1900 to 2000. More accurately, comparing the slope between 1907–1960 and 1960–2000, the gradient has become smaller (1.4 mm/year) in the latter period. In fact, the rise of sea level nearly stopped after 2005. This point will be discussed in Section 5.

Figure 4(b) shows changes of the global average temperature from 1890 to 2007 (from the Japan Meteorological Agency; the red line is added by the JMA.

Very similar figures have been published by NASA (GISS), NOAA, and others. Figure 4(b) shows also the amount of CO2 released in the atmosphere which began to rise rapidly in 1946. Although the global average temperature (T) changes can be approximated by a linear relation as a function of time (t) (T = at), CO2 changes are more like T =bt2.

Figure 4(c) presents an interpretation of Figure 4(b), showing temperature changes that consist of a linear change and ‘fluctuations’ superposed on it. The temperature record (thin blue line) is taken from the NOAA report 42 (see the insert in Figure 9), which is basically a smoothed version of the 5-year mean in Figure 4(b). The thick blue line from 1975 to 2000 will be discussed in Section 6.

In examining the linearity of temperature changes during the last century, Bryant noted that there are only a few points outside the 95% confidence limits of the linear approximation. The gradient of the straight line is about 0.5 °C/100 years.

There have been a number of discussions of the temperature during the LIA. It ranges from 0.5 C to 1.5 C below the present temperature. If we take it to be 1.0 C mainly on the basis of Figure 1a, the rate of increase during the last 200 years, namely between 1800 and 2000, is about 0.5 C/100 years.

3.2. Did the Recovery from the LIA End before 1900?

On the basis of the above studies, we find the recovery has continued to the present.

If we consider the present temperature is the normal temperature, then a casual inspection of Figures 1(a) and 1(b) might give an impression that the Earth has recovered from the LIA,.

However, in meteorology and climatology, it is not possible to define the absolute normal level (baseline) from which deviations (warming or cooling) can be measured or the end of the LIA can be determined. If we examine data over a longer period (say, 2000 and 10,000 years), we find the present temperature can be below the average temperature for chosen periods.

Figure 5 shows ice core temperatures at the GISP-2 site in Greenland for the past 2000 years and for the past 10,000 years. Clearly, the average temperature in the 20th century is not useful in answering our question in this subsection.

Furthermore, it is very important to note that recent temperature is in a rising trend. Thus, there is no basis to define the ending year of the LIA. It is more likely that the Earth is still in the process of recovery. What we have learned so far has significant implication for understanding the temperature rise in the 20th century. We will discuss this point in Section 6 (see Figure 9).

4. Possible Solar Causes

The cosmic-ray intensity from the year 1000 to the present shows solar activity was relatively low during the LIA and began to recover from about 1800.

The fact that the cosmic-ray intensity varied during the LIA suggests that non-terrestrial forces, more specifically solar activity, are involved in some components of climate change.

Figure 6 shows the solar modulation function of cosmic rays deduced from 10Be and 14C records from 1000 to 2000. When solar modulation is low, solar activity is low (but, cosmic-ray intensity is high). Conversely, when solar moderation is high, solar activity is high (but, the cosmic-ray intensity is low). Solar activity is represented by the sunspot number and its changes are well correlated with changes of the solar irradiance.

Therefore, Figure 6 represents the solar activity trend from 1000 to 2000, which may be compared with Figure 1(a). We see that solar activity was relatively low during the LIA and began to recover in about 1800. Therefore, we may speculate that solar irradiance is involved in causing the LIA and its recovery.

Changes of the solar irradiance during the sunspot cycle are rather small (1.3W/m2). However, the difference between the LIA period and the present may be a few times greater than 1.3W/m2. Therefore, although Nozawa et al. showed the solar effect on temperature changes during the 20th century was small, this subject requires much more detailed study with newer Global Climate Models by taking into account a prolonged period of a low solar irradiance (Scafetta and West), at least as its triggering effects.

Figure 1(a) shows the LIA began in about 1200 or 1300. Such a prolonged period of low solar irradiance may cause a significant climate change (cf. Scafetta and West). In this paper, we are mainly interested in possible causes of a long period change like the LIA. On the other hand, Soon examined the solar effect of a shorter period (130 years) and found a significant correlation with arctic temperature variations.

Figure 6 shows several minima: the Oort Minimum (1000-1100), the Wolf Minimum (1250-1350), the Spoer Minimum (1380-1510), and the Maunder Minimum (1620-1720). Intermittent increases of the solar modulation function (thus, low cosmic-ray intensities) were caused by a high solar activity.

Figure 1(a) suggests the LIA was not a continuously cool period. Since solar activity represented by the sunspot number correlates well with the solar irradiance, Figure 6 represents the general trend of changes of the solar irradiance among others.

5. Multi-Decadal Change

The purpose of this section is to explain why the warming has halted after 2000, despite the fact that we concluded in previous sections that the Earth is still in the recovery process from the LIA.

The multi-decadal oscillation may be halting the recovery from the LIA temporarily. This is similar to the situation from 1940 to 1975.

Figure 7 shows the global average temperature changes during the last several decades and the halting of warming in the last decade.

In Section 3, we suggested the prominent ‘fluctuations’ superposed on the linear recovery are the multi-decadal oscillations. Figure 4(c) shows a multi-decadal oscillation peaked in 1940, and the temperature decreased from 1940 to 1975 and then increased again after 1975 to 2000.

Thus, we speculate the situation in 2000 is similar to that in 1940. Using this assumption, we predict future temperature change will be flat or in a slightly declining trend during the next 30 years or so (see Section 6 and Figure 9). That is to say, the halting does not mean the end of the recovery from the LIA.

The halting after 2000 can also be observed in sea level change, a decrease of the heat content of the oceans and other factors.

Figure 8 shows the pattern of the Pacific Decadal Oscillation (PDO), which is a natural phenomenon. Top part shows the observed wind pattern over the Pacific Ocean. The middle part shows the PDO index. The bottom diagram is the same as Figure 4(c), reproduced here to show the striking resemblance of changes between PDO and the multi-decadal oscillation.

Although there is some phase difference between them, this similarity supports the inference that the fluctuations superposed on the linear change (the recovery from the LIA) are part the multi-decadal oscillation. The Pacific Ocean is large enough to contribute to the global average temperature. The Arctic Ocean shows a similar trend.

6. Summary

Here is a summary of what we have learned.

Figure 9 shows a large box which is the same as Figure 4(c). It suggests temperature changes from 1800 to 2000 can be explained mainly as a combination of the linear increase from about 1800-1850 and the multi-decadal oscillation. The blue line is taken from the NOAA data shown in a small box above the large box. The observed temperature in 2008 is shown by a red dot with a green arrow.

The IPCC assumes the temperature increase shown by the thick blue line was caused mostly by the human emissions of carbon dioxide, and predicts temperature will follow an extension of the blue line.

However, the meaning of the linearity of the recovery from 1800-1850 is crucial in understanding the cause of the warming in the last century. This linearity shows the temperature increase from 1850 to present is a natural recovery from the LIA and has not been caused by increases in CO2.

The rate of temperature increase since 1850 has not changed even though human emissions of CO2 in 2000 were at least 14 times greater than in 1900 and much greater than in 1850.

The simplest assumption is that the recovery from the LIA and the multi-decadal oscillation will continue during the next 100 years. This assumption allows us to predict the future trend to 2100. This trend is the natural trend and it is not affected by human activities.

7. Conclusions

  1. The Earth experienced the Little Ice Age (LIA) between 1200-1400 and 1800-1850. The temperature during the LIA was about 1C lower than present temperature. The solar irradiance was relatively low during the LIA.
  2. The gradual recovery from the 1800-1850 LIA was approximately linear. The recovery (warming) rate was about 0.5°C/100 years. This has important implications for understanding the present global warming.
  3. The same linear change continued from 1800-1850 to 2000. In this period, the solar irradiance began to recover from its low value during the LIA.
  4. The recovery from the LIA is still continuing today. There is no sign to indicate the recovery ended before 1900.
  5. The multi-decadal oscillation of 50 to 60 years is superposed on the linear change. It peaked in about 1940 and in 2000, causing the temporal halting of the recovery from the LIA.
  6. The negative trend after the peak in 1940 and 2000 overwhelmed the linear trend of the recovery, causing the cooling or halting of warming.
  7. These changes are natural changes, and in order to determine the contribution of the man-made greenhouse effect, there is a need to identify these natural changes correctly and accurately and remove them from the present global warming/cooling trend.
  8. The view presented predicts the temperature increase in 2100 will be 0.5C ± 0.2C, rather than 4C ± 2.0C predicted by the IPCC.


The Amount of Non-Fossil-Fuel CO2 in the Atmosphere Figure 3. (a) Retreat of glaciers in Glacier Bay, Alaska, began in 1800.